Allowance Allocation and CO2 intensity of the EU15 and ......Allowance Allocation and CO2 intensity...

Allowance Allocation and CO2intensity of the EU15 and

Norwegian refineries

Kristina Nilsson, Lars Zetterberg & Markus Åhman B1610

February, 2005

Report SummaryOrganisation/OrganizationIVL Svenska Miljöinstitutet ABIVL Swedish Environmental Research Institute Ltd. Projekttitel/Project title

Handel med utsläppsrätter – Olikafördelningssystem för utsläppsrätter och deraskonsekvenser – etapp 2

Adress/address

Box 210 60100 31 Stockholm Project sponsor

Preems Miljöstiftelse (The EnvironmentalFoundation of Preem),Naturvårdsverket (Swedish EnvironmentalProtection Agency)

Telefon/Telephone

+46 8 598 563 00

Rapportförfattare/author

Kristina Nilsson , Lars Zetterberg, Markus Åhman

Title and subtitle of the report

Allowance Allocation and CO2 intensity of the EU15 and Norwegian refineries

Sammanfattning/SummaryOn 1 January 2005 the European Union Emission Trading Scheme was launched. The launch waspreceded by an allocation process in each of the Member States. The main objective of this studywas to analyse the allocation in relation to CO2 efficiency for the mineral oil refining sector.

A CO2 intensity index for mineral oil refineries was defined and calculated for the refineries withinthe EU15 and Norway. The IVL CO2 intensity index is based both on the Solomon Energy IntensityIndex (EII), an assumed fuel mix and process-specific emissions. Due to uncertainties in input data,the determined values for the individual refineries are fairly uncertain, but the regional values canbe used to identify trends.

It was concluded that there are substantial differences in the CO2 intensity between refineries withindifferent regions/countries in the EU and these differences have not been considered in theallocation process. However, there seems to be a correlation between allocation and CO2 efficiencyfor refineries in different regions. With some exceptions countries where the mineral oil refiningindustry has a low CO2 intensity index have allocated relatively more than countries with industriesof high CO2 intensities.

Only a few countries have mentioned energy efficiency or reduction potential due to CO2 intensityof fuels used. Only one country (Denmark) has explicitly given a benchmark that will be used forallocation to new mineral oil refineries.Nyckelord samt ev. anknytning till geografiskt område eller näringsgren /Keywords

Climate efficiency, energy intensity, CO2 emissions, mineral oil refinery, EU ETS

Bibliografiska uppgifter/Bibliographic data

IVL Rapport/report B1610Rapporten beställs via /The report can be ordered via www.ivl.se, e-mail: [email protected], fax: 08-598 563 90 eller IVL, Box 210 60, 100 31Stockholm

Allowance Allocation and CO2 intensity of the EU15 and Norwegian refineries IVL Report B1610

4

AcknowledgementsThis research project was performed by IVL Swedish Environmental Research InstituteLtd. The Swedish Environmental Protection Agency and the Environmental Foundationof Preem financed the project. This report is part of the project “Handel medutsläppsrätter – Olika fördelningssystem för utsläppsrätter och deras konsekvenser –Etapp 2” in which two other summarising reports have been written, Zetterberg et al(2004) and Åhman (2004). Zetterberg et al’s (2004) report, “Analysis of nationalallocation plans for the EU ETS”, is particularly closely related to this report. There isalso a report for the first phase of the project written by Åhman and Zetterberg (2004),“Options for Emission Allowance Allocation under the EU Emissions TradingDirective”.


6

AbstractOn 1 January 2005, the European Union Emission Trading Scheme (EU ETS) waslaunched. The launch has been preceded by an allocation process in each of the MemberStates. The main objective of this study was to analyse the allocation in relation to CO2

efficiency for the mineral oil refining sector.

A CO2 intensity index for mineral oil refineries has been defined and calculated for therefineries within the EU15 and Norway. The IVL CO2 intensity index is based both onthe Solomon Energy Intensity Index (EII), an assumed fuel mix, and process-specificemissions. Due to uncertainties in input data, the determined values for the individualrefineries are quite uncertain. However, the regional values can be used to identifytrends.

It was concluded that there are substantial differences in the CO2 intensity betweenrefineries within different regions/countries in the EU and these differences have notbeen considered in the allocation process. Only a few countries have mentioned energyefficiency or reduction potential due to CO2 intensity of fuels used. Only one country(Denmark) has explicitly given a benchmark that will be used for allocation to newmineral oil refineries.

The allocation has generally been based on historic emissions, which will result inrefineries with historically higher emissions being allocated larger amounts thanrefineries with historically lower emissions. This might be favourable for refineries thatrecently have performed emission-reducing measures but might be less favourable forrefineries that during a long time period have implemented emission-reducing measures.


7

Table of Contents

Acknowledgements .......................................................................................................... 4

Abstract............................................................................................................................. 6

1 Introduction ............................................................................................................... 9

2 Objective ................................................................................................................... 9

3 CO2 intensity for mineral oil refineries ................................................................... 103.1 Sources of CO2 emissions at mineral oil refineries ............................................. 10

3.1.1 Combustion sources – stationary devices................................................... 113.1.2 Point sources – process vents ..................................................................... 123.1.3 Point sources - other vents.......................................................................... 153.1.4 Non-point sources....................................................................................... 163.1.5 Indirects ...................................................................................................... 16

3.2 General theory of CO2 intensity .......................................................................... 163.2.1 The Solomon energy intensity index .......................................................... 163.2.2 The IVL CO2 intensity index for fuel refineries - theory ........................... 17

3.2.2.1 CO2 emissions considered in the IVL CO2 intensity index..................... 183.2.2.2 Energy-related CO2 emissions ................................................................ 183.2.2.3 Process-related CO2 emissions................................................................ 193.2.2.4 The total refinery CO2 intensity............................................................... 203.2.2.5 Assumptions of the IVL CO2 intensity index for fuel refineries............. 20

4 Different tiers for calculating the IVL CO2 intensity index .................................... 234.3 Tier 1. Calculation of the CO2 intensity index – data from questionnaires......... 24

4.3.1 Energy-related CO2 emissions – combustion of fossil fuels ...................... 254.3.2 Process vents – process-related CO2 emissions.......................................... 26

4.3.2.1 Non-replaceable fuels.............................................................................. 264.3.2.2 Process vents ........................................................................................... 26

4.3.3 Actual CO2 emissions................................................................................. 274.4 Tier 2. Calculation of the CO2 intensity – O&GJ /Solomon ............................... 27

4.4.1 Energy related emissions – combustion of fossil fuels............................... 284.4.2 Process vents – process-specific CO2 emissions ........................................ 294.4.3 Actual emissions......................................................................................... 29

4.5 Tier 3. Calculation of the CO2 intensity – other data from producers................. 30

5 Data collection ........................................................................................................ 305.1 Tier 1.................................................................................................................... 315.2 Tier 2.................................................................................................................... 315.3 Tier 3.................................................................................................................... 32

6 Uncertainties ........................................................................................................... 32


8

6.1 Tier 1.................................................................................................................... 326.2 Tier 2.................................................................................................................... 326.3 Tier 3.................................................................................................................... 34

7 Results ..................................................................................................................... 347.1 The CO2 intensity for refineries by country ........................................................ 34

7.1.1 Austria ........................................................................................................ 347.1.2 Belgium ...................................................................................................... 357.1.3 Denmark ..................................................................................................... 357.1.4 Finland ........................................................................................................ 367.1.5 France ......................................................................................................... 367.1.6 Germany ..................................................................................................... 377.1.7 Greece......................................................................................................... 387.1.8 Ireland......................................................................................................... 397.1.9 Italy............................................................................................................. 397.1.10 The Netherlands.......................................................................................... 407.1.11 Norway ....................................................................................................... 417.1.12 Portugal....................................................................................................... 417.1.13 Spain ........................................................................................................... 417.1.14 Sweden........................................................................................................ 427.1.15 The United Kingdom.................................................................................. 42

7.2 Summary of the IVL CO2 intensity index ........................................................... 437.3 Allocation to the Mineral Oil Refining Industry ................................................. 44

8 Discussion ............................................................................................................... 498.1 The differences in CO2 intensity between refineries ........................................... 498.2 For what purposes can the IVL CO2 intensity index be used? ............................ 528.3 Expectations and Outcome of the EU ETS national allocation plans ................. 528.4 Future challenges................................................................................................. 53

9 Conclusions ............................................................................................................. 55

10 Further research/improvements............................................................................... 56

11 References ............................................................................................................... 5711.1 Literature.......................................................................................................... 5711.2 Personal communications ................................................................................ 5911.3 Web-sites.......................................................................................................... 59

Appendix 1.................................................................................................................. 60


9

1 IntroductionOn 1 January 2005, the European Union launched an emission-trading scheme coveringcarbon dioxide emissions from the energy and some of the industry sectors (Directive2003/87EC). Prior to the start of the trading scheme, every Member State was obliged tosubmit a national allocation plan describing how the initial amount of allowances wouldbe allocated among the covered installations. The Commission, which was the authorityto approve or reject the allocation plans, had given a list of criteria to be followed andguidelines to support the design of the national allocation plans. In the approval process,no harmonising between countries or sectors has been done other than making sure theprinciples do not violate any of the criteria listed in Annex III to the Directive2003/87/EC (Zetterberg et al, 2004).

The European Union Emission Trading Scheme (EU ETS) approves of agreements withnon-EU countries in order to provide for mutual recognition of emission allowancesbetween the Community scheme and other emission allowance trading schemes.Norway has showed interest in such an agreement and since Norwegian refineries arecompetitors to not only the Swedish refineries but to most refineries within the EU, theNorwegian refineries were also included in this study.

2 ObjectiveThe objective of this study was to:

• Quantify CO2 emissions, production and “climate efficiency” for a selection of EUrefineries. This included developing a new measure of the “climate efficiency” forfuel refineries.

• Assess the allocation methodologies and actual allocation for a selection of the EUrefineries.

• Analyse the allocation in relation to the climate efficiency. Do refineries with lowspecific emissions benefit from the allocation methodology used or are theconditions reversed? How is the allocation correlated to emissions? How is theallocation correlated to the production? To determine how the allocations are relatedto the potential of reducing emissions and if the presence of energy-efficienttechnologies is considered when making the allocation.

The objective of this study was not to point out the exact differences betweeninstallations. However, the analysis points out differences in allocations to the oilrefining sectors in different Member States. The aim was mainly to identify generaltrends in order to discover disadvantages for certain installations / types of installations


10

or the whole sector in a country and thereby prevent those disadvantages from beingoverlooked in the national allocation plans, which would create biases in thecompetitive situation.

This report will mainly focus on the quantification of CO2 emissions and CO2 efficiencybut will, to a certain extent also, analyse the results of the allocation process for themineral oil refineries. A more complete assessment of the allocation methodologies hasalready been done in Zetterberg et al (2004).

3 CO2 intensity for mineral oil refineriesThe main objectives of this study were first, to define a measure of climate efficiencyfor mineral oil refineries in order to see if this has been considered in the allocationprocess and then, to assess if highly climate-efficient refineries have been treated morefavourably than less CO2-efficient refineries. Since only carbon dioxide is included inthe EU ETS during the first period (2005-2007), it was decided that ‘climate efficiency’would only reflect the CO2 efficiency and not efficiency with respect to othergreenhouse gases such as CH4 or N2O. For practical reasons, an intensity index insteadof an efficiency index was determined. A description of the developed CO2 intensityindex for mineral oil refineries is given below.

3.1 Sources of CO2 emissions at mineral oil refineries

The CO2 emissions from a mineral oil refinery are affected by many sources:

• The complexity of the refinery (number of different processes)

• Process-related fuels that have to be burned

• Quality of product slate delivered (e.g., low sulphur fuels)

• Quality of crudes and other raw materials used in the refining process

Shires and Loughran (2004) have, on the behalf of the American Petroleum Institute,put together a compendium on methodologies on how to estimate greenhouse gasemissions from the oil and gas industry. According to this compendium, the CO2

emissions from mineral oil refineries can be divided into the following sourcecategories:

Combustion sources – Stationary Devices– Boilers, process heaters, turbines, engines, flares, catalytic and thermal oxidisers, cokecalcining kilns, incinerators.Point sources – Process vents– catalytic cracking, catalytic reforming, catalyst regeneration, thermal cracking,coking, hydrogen production, sulphur recovery units, asphalt production.


11

Point sources - other venting– Storage tanks, pneumatic devices, loading racksNon-point sources – fugitive emissions– Fuel gas system leaks, other process equipment leaksNon-point sources – other non-point sources– Waste water collection and treating, sludge/solids handling, cooling towers.Non-routine activities – other releases– Pressure relief valves (PRV), emergency shut downs (ESD).Indirects– Electricity usage/production, steam generation/import

In this study the sources of CO2 emissions from refineries were divided into fivecategories:

• combustion sources

• point sources - process vents

• point sources - other vents

• non-point sources and

• indirects

According to the Guidelines for monitoring and reporting of greenhouse gas emissionsfrom installations included in the EU ETS (2004/156/EC), the monitoring of greenhousegas emissions from a mineral oil refinery shall include all emissions from combustionand production processes as occurring in refineries. In the guidelines, a list of potentialsources of CO2 emissions is given. The sources are divided into two categories:

1. Energy-related combustion:- Boilers, Process heaters / treaters, Internal combustion engines / turbines, Catalyticand thermal oxidisers, Coke calcining kilns, Emergency/ standby generators, Flares,Incinerators and Crackers.

2. Processes:- Hydrogen production installations, Catalytic regeneration (from catalytic crackingand other catalytic processes) and Cokers (flexi-coking, delayed coking).

Below is a closer description of the categories used in this study.

3.1.1 Combustion sources – stationary devices

The main source of CO2 emissions in refineries is the combustion of fuels forproduction of steam, heat and electricity. Most of the processes in a refinery requireheating and, therefore, combustion of fuels is necessary. The operation of the refineryalso requires some use of electricity and many processes are dependent on steam access.


12

Shires and Loughran (2004) also mention flaring as a part of this source category.Flaring is a process where surplus gas is burned. This combustion is not input to anyother process but is only performed in order to eliminate surplus gas. Emissionsresulting from combustion in transport devices (i.e., vehicles, ships, etc.) should also beincluded in this category (Shires & Loughran, 2004). Comparing the definition made byShires and Loughran (2004) and the European Commission in 2004/156/EC, one canconclude that the combustion source categories are quite similar. The Commission doesnot, however, include emissions resulting from combustion in transport devices such asvehicles and ships. The Commission includes emergency generators, which Shires andLoughran (2004) included in non-point sources.

3.1.2 Point sources – process vents

According to Shires and Loughran (2004), there are a few different processes that giverise to CO2 emissions other than the combustion of fossil fuels as an energy source forthe process. These processes are:

• Hydrogen production

• Regeneration of catalytic cracker catalyst and regeneration of other catalysts

• Cokers

• Other process vents

The potential sources of process-related CO2 emissions at mineral oil refineries listed bythe European Commission (2004/156/EC) are the same (see the beginning of thischapter).

Hydrogen productionThe hydrogen plant produces a significant quantity of CO2, which may be furtherprocessed for other uses or may be vented to the atmosphere. The quantity of CO2

vented depends on the carbon to hydrogen rate of the feed gas. Most refineries use aprocess (steam reforming) where H2 is produced from CH4, but there are some plantsthat operate with other feedstock gases (European Commission, 2003). The chemicalreaction can be expressed by Equation 3.1Equation 3.1. Chemical reaction in hydrogen production plant.

222)22( )13(2 xCOHxOxHHC xx ++→++

According to Shires and Loughran (2004), there is a simple approach to estimate theCO2 vent rate based on the average feed gas composition. This approach should beadequate for most refineries where the feed gas is similar to natural gas (i.e.,predominantly CH4 with small percentages of other low molecular weight


13

hydrocarbons). The approach is based on an emission factor of 13.411 tonnes CO2 permillion scf 2of H2 produced. Some refineries use refinery gas, LPG or light naphtha asfeed gas to the hydrogen production unit. This will result in higher specific CO2

emissions (emissions per produced ton of hydrogen) than when using natural gas (due toa higher C to H ratio). Note that the CO2 vent stream described by the reaction abovedoes not include CO2 emissions from process heaters associated with the H2 plant(Shires & Loughran, 2004). Those emissions should be treated like other combustionsources.

There are some refineries that use another process for producing H2, namely the partialoxidation process. In the case of such a process, site-specific data or an engineeringapproach should be used in order to estimate the CO2 emissions. However, if no suchdata is available, Shires and Loughran (2004) suggest that a conservative approachassuming full conversion and using the simple approach emission factor be used.

According to the guidelines given by the European Commission (2004/156/EC), theCO2 emissions resulting from hydrogen production should be based on the followingequation:

CO2-emissions = activity datainput * emission factor , whereactivity datainput = amount of hydrocarbon feed [t feed]emission factor = either reference value of 2.9 t CO2/t feed processed3 or valuecalculated from the carbon content of the feed gas [t CO2/t feed].

Catalytic cracker Catalytic cracking is the most widely used conversion process for upgrading heavierhydrocarbons into more valuable lower boiling hydrocarbons. Normally, the main feedstream to the process is the vacuum distillate stream from the vacuum distillation unit.This process uses heat and a catalyst to break the larger molecules into smaller ones(European Commission, 2003).In the catalytic cracking process, coke is deposited on the catalyst as a by-product of thereaction. That coke is burned in order to restore the activity of the catalyst. The coke iscontinuously burned off in the regenerator. This process is a significant source of CO2-emissions. According to Shires and Loughran (2004), there are two common approachesthat can be used to estimate the CO2 emissions from catalytic cracker catalystregeneration:

1 This emission factor has been determined based on a feed gas with the following composition: CH4 =93.07 % (vol.), C2H6 = 3.21% (vol.), C3H8 = 0.59 % (vol.), higher hydrocarbons = 0.32% (vol.) and Non-hydrocarbons = 2.81% (vol.) (Shires & Loughran, 2004).2 Scf = standard cubic feet.3 Conservatively based on ethane.


14

• process calculations based on coke burned

• process calculations based on air rates and CO/CO2 concentrations

Equation 3.2 describes how the emissions from the regeneration of the catalyst aredetermined by using the first approach.

Equation 3.2 . CO2 emissions from regeneration of catalytic cracker catalyst.

1244

2⋅⋅= CFCCECO

Where: 2COE = emission of CO2/year

CC = coke burn rate in tonnes/yearCF = fraction of carbon in the coke burned per weight44/12 = conversion factor from CO2- C to CO2

The approach suggested by the European Commission (2004/156/EC) is also describedby Equation 3.2, where the carbon fraction in the coke burned is determined inaccordance with the provisions of section 10 in Annex 1 of the Commission Decision(2004/156/EC) 4. The Commission suggests that the same method should also be usedalso for determination of CO2 emissions from other catalyst regeneration.

Other catalyst regenerationThere are also other process units at a refinery with catalysts that have to beregenerated. Examples of such processes are catalytic reformers and hydro-processingunits. Hydro-processing units are used for desulphurisation and conversion of anyfractions into products with a lower molecular weight than the feed. Normallyhydrogen, heat and catalysts are used in the hydro-processing units. In catalyticreformers, the heavy naphtha leaving the hydrotreating unit is upgraded for use as agasoline blend stock. The octane number is significantly improved in the catalyticreforming process. Both hydro-processing and catalytic reforming units do havecatalysts that have to be regenerated in order to maintain the function of the catalyst.There are two main types of catalyst regeneration:

• continuous regeneration

• intermittent regeneration: cyclic or semi-regenerative

The emissions from these types of catalyst regeneration can be determined by usingEquation 3.2.

4 The specific procedure to determine process specific emission factors should follow CEN, ISO ornational standards. If no such standards are available, procedures can be carried out in accordance withdraft standards or industry best practice guidelines. (For further information, see Commission Decision of29 Jan. 2004).


15

CokersCoking is a severe thermal cracking process used mainly to reduce refinery productionof low-value residual fuel oils and transform them into transportation fuels, such asgasoline and diesel. As part of the process, coking also produces petroleum coke, whichis essentially solid carbon with varying amounts of impurities.There are three main types of cokers:

• Delayed cokers

• Fluid cokers

• Flexi-cokers

Only fluid cokers and flexi-cokers have CO2 emissions resulting from the coke burner.Delayed cokers do not have CO2 emissions, other than from their process heaters thatare calculated as any other combustion source (Shires & Loughran, 2004). Fluid cokersand flexi-cokers have a vent resulting from the coke burner. In this study, the CO2

emissions from the coke burner were estimated by assuming that all of the carbon in thecoke is oxidised to CO2. Equation 3.2 for catalytic cracking units can also be used forfluid cokers or flexi-cokers.According to the Commission (2004/156/EC), the CO2 emissions from cokers should becalculated by Equation 3.2 where the carbon fraction [t C/t coke] is based on industrybest practice guidelines.

Other process ventsThere are also other processes at a refinery that could be considered as potential sourcesof CO2 emissions, such as:

• Asphalt blowing

• Thermal cracking

• Sulphur recovery units

In Shires and Loughran (2004), these sources are considered to be negligible incomparison to other sources. There are no specific guidelines on how to calculate theseprocess vents given by the European Commission (2004/156/EC).

3.1.3 Point sources - other vents

These emissions constitute vents from sources such as storage tanks and pneumaticdevices. According to Shires and Loughran (2004), the emissions from this source aremainly CH4, whereas the CO2 emissions are small. There are no specific guidelines onhow to calculate these process vents given by the European Commission 2004/156/EC.


16

3.1.4 Non-point sources

According to Shires and Loughran (2004), non-point sources are negligible incomparison to the other sources. Non-routine activities are also, in most cases,negligible (mostly because they do not occur; if they do occur, they could be ofsignificance). In the guidelines given by the European Commission (2004/156/EC), thenon-routine activities are included in the combustion sources.

3.1.5 Indirects

Indirect sources include the balance of imported electricity, heat and steam. The mainobjective of this study was to compare the allocation and allocation methodology forrefineries in different EU Member States. The requirement of CO2 emission allowancesis related to the actual emissions at the refinery. No allowances will be required foremissions that have occurred or will occur at other places than at the refinery. Thismeans that the refinery is not required to hold emission allowances for CO2 emissionsrelated to electricity, steam or heat generated at an external unit but consumed at therefinery. On the other hand, the refinery is required to hold emission allowances for anyCO2 emissions generated during the production of heat, steam and electricity at the site,even if some of the surplus is exported from the refinery for use elsewhere.

3.2 General theory of CO2 intensity

As described in the previous section, there are different sources of CO2 emissions froma refinery. However, the major part of the emissions is connected to the combustion offossil fuels, in other words the energy demand. The CO2 intensity is therefore closelyrelated to the energy intensity.

3.2.1 The Solomon energy intensity index

There is today a concept for comparing the energy intensity5 of mineral oil refineries –the Solomon Energy Intensity Index (EII). Solomon Associates determine the energyintensity index for refineries by the following equation:Equation 3.3 The Solomon Energy Intensity Index

100⋅=nconsumptioenergyreference

nconsumptioenergyactualEII

The actual energy consumption is determined by adding purchased energy (such aselectricity and steam) to used fuels and subtracting sold energy (such as electricity, heat

5 Intensity = 1/efficiency.


17

or steam). The value of primary energy is used, i.e. purchased electricity is converted toprimary energy by assuming 9.59 MJ/kWh. The reference energy consumption isdetermined by the following equation:Equation 3.4 Reference energy requirements according to Solomon

∑ = ⋅= mi i

ei

Qreference

E 1 )(

where Qi = actual load /production in process i,ei = the Solomon process unit energy standard (either just a constant or an equation withprocess-related variables) for process i,m = the total number of different processes at the refinery

The Solomon process unit energy standards are individual expressions for each of theprocesses in the refinery and state the average standard energy consumption for theprocess in comparison to process load or output (depending on the process). SolomonAssociates have determined the values of the process unit energy standards by practicalevaluation of 300 refineries worldwide. The currently used process unit energystandards are based on average industry efficiency for each process unit types in themid-1980’s (personal comm. Trout, 2004). At that time, the industry average EII wasclose to 100, whereas the index today average 90 or less. The reason for the decrease inthe index over the years is that the refineries have improved their energy efficiency.Note, however, that the EII is not designed to determine CO2 performance.

3.2.2 The IVL CO2 intensity index for fuel refineries - theory

IVL has defined a CO2 intensity index for fuel refineries, which will be determined in away similar to the Solomon energy intensity index. The CO2 intensity will be a relativemeasure of the CO2 performance of a refinery in comparison to a sector expected value.The actual emissions of a production unit, a refinery, will be compared to a sectorexpected value. Generally, the CO2 intensity can be described by Equation 3.5.Equation 3.5. The CO2 intensity

100,2

,2,2

⋅=R

COA

CO

intensityCO

Where CO2, A is the actual emissions from a specific refinery and CO2,R is the referenceCO2 emissions from a reference refinery of the same size and complexity. Note that theindex is just a relative measure and whether or not the value will be above or below 100depends on if the reference emissions are determined according to best technology, theaverage value (such as for the Solomon EII) or something in-between. If the referenceemissions were determined according to the use of best technology, all refineries wouldhave a value equal to or above 100. On the other hand, if the reference emissions were


18

determined according to average emissions, half of the refineries would have a CO2

intensity index below 100. In the IVL CO2 intensity index, the reference emissions havebeen determined at a level above the best available technology but below the averagevalue.

3.2.2.1 CO2 emissions considered in the IVL CO2 intensity index

As described in section 3.1, there are many sources of CO2 emissions at a refinery. Forthe purpose of describing the CO2 intensity, the CO2 emissions from each process weredivided into two sub-groups (the same sub-groups as used in the Commission Decision2004/156/EC):

• energy (or fuel) related emissions

• process-related emissions (raw-material dependent)

In this study, the energy-related emissions were defined as the emissions due tocombustion of fossil fuels (and the use of electricity). This includes all types of fuels,both replaceable and non-replaceable (such as coke in a catalytic cracker). The process-related emissions are emissions not related to combustion of fossil fuel with the purposeto produce energy. One such example is the emissions from a hydrogen production plantwhere there are chemical reactions not related to the oxidation of a fossil fuel, that occurin the process and give rise to CO2 emissions.

3.2.2.2 Energy-related CO2 emissions

The energy-related reference emissions are described by Equation 3.6 below.Equation 3.6. The reference energy-related CO2 emissions for a specific process i, in refinery B.

iSnconsumptioenergy

iSCO

iSproduction

iSnconsumptioenergy

iBproduction

iRECO

,

,,2

,

,,,,2

⋅⋅=

The denotation R indicates that it is the reference emissions, the E indicates that it isenergy-related emissions, B indicates a specific refinery, S indicates the sectorreference6 value set for the mineral oil refining sector and i denotes that it is a process-specific value. The production is the actual production at the refinery. The sectoraverage energy requirements for a specific process is described by the second factor:

=

feedbblKBtu

iprocessforstandardenergyunitprocessSolomoniSproduction

iSnconsumptioenergy

,

,

6 In this study we have used the emissions factor of refinery gas as a reference value for the use of fuels(except for non-replaceable fuels).


19

where the Solomon process unit energy standards have been determined by SolomonAssociates through the evaluation of over 300 refineries worldwide. The third factor on

the right hand side of Equation 3.6, iS

nconsumptioenergyiS

CO

,

,;2 , describes the emissions

related to fuel use for the specific process. Equation 3.7 describes how the fuel-relatedCO2 emission from a specific process is determined by using emission factors (gCO2/MJ). These emission factors are determined by assuming a fuel mix for eachprocess.Equation 3.7. The reference CO2 emissions for a specific process.

⋅= ∑ KBtu

COtfactoremissionnenergyCO

jjj

iS

iS 2

,

,,2

Where iS,

j,S,

totalsupplyenergyfuelfromsupplyenergy in=

Hence, the use of fuels (j) in each process (i) is determined in order to estimate theexpected energy-related CO2 emissions from each process.

3.2.2.3 Process-related CO2 emissions

The only identified process vent was the hydrogen plant. Often, the catalytic crackerand other processes where coke is formed and has to be burnt off are mentioned asprocess-specific emissions. Since the coke formed on catalysts and burnt in the processcontributes to the energy supply of the process, these emissions were treated in thisstudy as non-replaceable fuels under energy-related emissions.Equation 3.8. The reference process-specific CO2 emissions of process i.

iSPproduction

iSPCO

iBproduction

iRPCO

,

,,2,,,2⋅=

where R indicates that it is the reference emissions and P that the emissions are processrelated. The other denotations have occurred in earlier equations. The second factor onthe right hand side of Equation 3.8 is the sector average process related CO2 emissionfactor for the process i:

productt

COtproduction

CO

iSP

iSP 2

,

,,2 .


20

3.2.2.4 The total refinery CO2 intensity

The reference emissions for each process of the refinery were described by summarisingthe energy-related emissions and the process-specific emissions of that specific process.Equation 3.9. The total reference CO2 emissions for a process i.

iRPCO

iRECO

iRCO

,,2,,2,,2+=

The R stands for the reference emissions and E indicates that it is the energy-relatedemissions (due to combustion of fossil fuels) that are considered. P indicates process-specific emissions. The actual CO2 emissions of a process are the sum of the actualenergy-related emissions and the actual process-related emissions.Equation 3.10. The total actual CO2 emissions for a process i.

iAPCO

iAECO

iACO

,,2,,2,,2+=

The refinery CO2 intensity was determined by the following equationEquation 3.11. The CO2 intensity of a refinery with i process units.

∑ +

∑ += m

i iRPCOiRECO

m

i iAPCOiAECO

intensityCO

,,2,,2

,,2,,2,2

The same denotation as in Equation 3.6 is used and the m stands for the total number ofprocesses at the refinery.

3.2.2.5 Assumptions of the IVL CO2 intensity index for fuel refineries

The IVL CO2 intensity index for mineral oil refineries was based on the assumptionsdescribed in the following sections.

General assumption

It was assumed that energy demands for Utilities, Losses and Off-sites (one of theSolomon process unit energy standards) could be left out when calculating the referenceCO2 emissions. This includes energy consumed in utility distribution systems andoperation of product blending, tank farms and environmental facilities. Some of theenergy is demand of electricity for machines (e.g., motor driven pumps, compressors),some of it is heat loss during transportation (which again most probably could beminimised to negligible levels by designing the refinery in an efficient way and usingheat exchangers), and some of it might be from other sources such as vents from storageor waste water plants. A reasonable assumption is that leaving these energy


21

consumption sources out will result in an equal offset for all refineries. However, in theSolomon process unit energy standard for Utilities Losses and Offsites, the compositerefinery configuration is one of the variables. This means that the complexity of therefinery influences the amount of energy needed for these sources. The offset wouldtherefore not be equal for all refineries, but for refineries with similar compositerefinery configuration. Since this is the only factor that considers electricity production,special consideration has been given to refineries that sell electricity to external units.This will reduce the differences between refineries of different complexity.

Combustion sources – stationary devicesSince flaring is a costly waste of energy, this process is limited as much as possible inall refineries. Therefore, no consideration to flaring was taken when calculating thereference emissions. The emissions, however, were still included in the actual totalemissions.

Emissions from transport devices were considered negligible in comparison to othersources and were therefore not included when determining the reference CO2 emissions.It should also be noted that CO2 emissions from transport devices are not included in theEU emission trading system.

Point sources – process ventsThe reference emissions from hydrogen production plants were determined by usingthe simple approach emission factor described in section 3.1.2. The simple approachemission factor was used both for steam reforming processes and partial oxidationprocesses. The reason for not using the suggested method in the Commission Decision2004/156/EC as described in section 3.1.2 was the availability of data. Output data (i.e.,amount of produced hydrogen) was more easily accessible than amount of hydrocarbonfed to the process.

The reference emissions resulting from the regeneration of the catalytic crackercatalyst and other catalysts were determined by using Equation 3.2, which is a methodboth suggested by the European Commission (2004/156/EC) and Shires and Loughran(2004). The reason for using this approach was also the availability of data; thisapproach is the least data intensive. The only process where catalyst regeneration wasconsidered was the fluid catalytic cracker unit. All other catalyst regeneration wasconsidered to be negligible in comparison. An example is given in the box below. Itshould be noted that the example given in Shires and Loughran (2004) for continuousregeneration uses a very high value of catalyst re-circulation (personal comm. Brinck2004, Magill, 2004). CO2 emissions from intermittent regeneration were considered tobe negligible in this study (and in Shires and Loughran, 2004) in comparison to othersources and were therefore left out when determining the reference CO2 emissions. Theemissions are, however, included in the total actual emissions.


22

Example: A FCC (Fluid Catalytic Cracker) unit is of the capacity 25,000 bbl/cd [barrelsper calendar day]

Data from Solomon study:UUOT (unit utilisation outside turnarounds) = 0.88Fresh feed density = 910.6 kg/m3

Coke on catalyst (yield vol. % of fresh feed) = 4.8%Conversion factor 1 bbl = 0.159 m3

Coke burn rate per year:

yrton

yrkgbblmcdcdtonmkgbbl

700,53365001.0153.06.910048.088.0000,253

3=

⋅⋅⋅⋅⋅⋅

CO2 emissions from FCC according to Equation 3.2:

yrCOton /900,1964412

1700,532=⋅

⋅

The emissions could also be determined based on coke burn rate estimated by theoperator.

The reference CO2 emissions from fluid cokers and flexi-cokers should have beendetermined by Equation 3.2. However, due to lack of process –data, this could not bedone. Only one refinery in the study had a fluid coker and one other had a flexi-coker.The amount of coke burnt could therefore not be estimated, which led to anunderestimate of the reference CO2 emissions connected to non-replaceable fuels forthose two refineries.

Note that all coke burnt was considered as non-replaceable fuel. This means that theenergy content in the coke burnt was considered to account for a part of the energydemand of the specific process.

Point sources – other ventsNo CO2 emissions from this source category were considered when determining thereference CO2 emissions, mainly because the CO2 emissions from these sources couldbe considered negligible compared to other sources. However, if a climate intensityindex were to be calculated including greenhouse gases other than carbon dioxide, thisassumption might have to be reconsidered since the main emissions from these sourcesare methane.

Non-point sourcesThese sources were not included in the IVL CO2 intensity index since these sources canbe considered negligible compared to other sources (see section 3.1.4).


23

IndirectsThis study focused on emissions that actually occur at the refineries. No adjustments ofreference CO2 emissions were made to compensate for purchased steam or electricity orsold heat or steam. The consequence of that is that if a refinery purchases a lot of itselectricity or steam from external producers, the reference CO2 emissions will beoverestimates. The operator will have lower emissions due to the fact that the emissionsemanating from the production of the electricity or steam purchased occurred elsewhere.This refinery will therefore appear to have a low CO2 intensity. Compensation wasmade for sold electricity. An extra amount of energy demand (and hence CO2

emissions) was added for the amount of sold electricity. The reason for consideringexported electricity is that this can be seen as a separate product. Of course the sameargument could be used for heat export, but in this study, heat delivery to external userswas seen as a way of efficiently using excess heat and not as a source for extra fuelrequirements at the refinery. In order to generate extra electricity, fuel is needed andextra fuel consumption is considered in the index calculations.

4 Different tiers for calculating the IVL CO2intensity index

The original approach for determining the IVL CO2 intensity index of the refinerieswithin Norway and EU15 was to ask for the necessary data via a questionnaire sent toeach of the refineries. However, very few of the 97 questionnaires were returned withfull information. Many of the refineries were in the midst of the process of deliveringdata on historic emissions to their governments and had difficulty, either due to lack oftime or sensitive information, providing all the data asked for in the questionnaire. Dueto this reason, we had to find other sources in order to get the data required forcalculating the CO2 intensity index. Three different tiers of determining the CO2

intensity index were developed, requiring different amounts of refinery-specific data.

Tier 1. The refineries delivered data by answering the questionnaire. Data givenincluded: Solomon EII-index, fuel mix, amount of purchased electricity, CO2 emissionsfrom processes with process-specific CO2 emissions, and actual CO2 emissions. Thedata given was valid for 2002. The values on process capacities given in Stell (2002)were also updated by the operators, along with values given on actual utilisation of eachprocess.

Tier 2. In the case where no answer to the questionnaire was received, the informationof refinery configuration (process composition) and production capacity given in the Oiland Gas Journal’s (O&GJ) 2001 worldwide refinery survey (O&GJ, 2001) was usedtogether with statistical data on unit utilisation and other operational statistics from theSolomon study to determine the reference emissions. The actual CO2 emissions fromthe EU15 refineries are available in the EPER database (http://eper.eea.eu.int/eper/). Alldata used in this tier was valid for 2001.


24

Tier 3. In some cases environmental reports from the refineries contained the neededdata to determine the reference emissions and the actual emissions.

The following sections describe which data and calculation methodologies were used tocalculate the CO2 intensities for the refineries according to the different tiers.

4.3 Tier 1. Calculation of the CO2 intensity index – data fromquestionnaires

The tier 1 methodology was used for refineries that answered the questionnaire7. Sinceall data asked for in the questionnaire was for 2002, the CO2 intensity indices calculatedaccording to this tier are valid for 2002. In the questionnaires, the amount of combustedfuels for running the processes, including the flared volume, was given. The amount ofpurchased electricity was also given.

In order to calculate the reference CO2 emissions, the reference energy consumptionaccording to Solomon was first determined. The Solomon energy efficiency index,which is calculated by Equation 3.3, was used to determine the reference energyconsumption. First, the actual energy consumption was determined by the followingequation:Equation 4.1. Actual energy consumption at refinery.

balanceheatbalancesteambalanceyelectricitvaluethermalfuelusagefuelnconsumptioenergyactual +++⋅=

Using the information available from the questionnaires the Solomon reference energyconsumption was then determined by combining Equation 3.3 and Equation 4.1 toEquation 4.2.

Equation 4.2 The Solomon reference energy consumption according to tier 1.

100/)(

EIISolomonbalanceheatbalancesteambalanceyelectricitvaluethermalfuelusagefuelnconsumptioenergyReference +++⋅

=

The energy output from the burned fuels was calculated by simply multiplying fuelconsumption by fuel thermal value. The electricity balance is the balance of sold andpurchased electricity. The balance is a positive value if the refinery is a net importer ofelectricity and is negative if the refinery is a net seller of electricity. The steam and heatbalances are analogous to the electricity balance. The total amount of energy achievedwas divided by the Solomon EII /100 for the refinery, which gave us the reference

7 Note: Only five refineries filled out the questionnaire, completely or partially.


25

energy needed according to Solomon. As mentioned in Section 3.2.2.5, a fewadjustments were made to the Solomon approach and the reference energy required forutilities, losses and offsites was subtracted from the total amount of reference energyneeded. This resulted in an adjusted reference energy amount that we refer to as the‘IVL reference energy consumption’. The reference CO2 emissions were determinedaccording to Equation 4.3.

Equation 4.3 Reference CO2 emissions due to combustion of fossil fuels

emissionsCOrelatedprocess)i

factoremissioni

fueli

of(fractionnconsumptioenergyreference

emissionsCOReference

2

2

+⋅∑⋅

=

The reference CO2 emissions were calculated by using the reference energy, a fixedassumed fuel mix (equal to all refineries) and emission factors.

Note that the guidelines from the European Commission (2004/156/EC) state that whendetermining the CO2 emissions from combustion of fossil fuels, one should use valuesof fuel consumption, emission factors and oxidation factors. The default values ofoxidation factors are 0.99 for solid fuels and 0.995 for gaseous fuels. In this study, wedid not consider the oxidation factors when determining the reference CO2 emissions.

4.3.1 Energy-related CO2 emissions – combustion of fossil fuels

Both the combustion of non-replaceable fuels (process-related fuels) and thecombustion of other fuels was, in this study, considered as energy related emissions.However, in this study the combustion of non-replaceable fuels was considered asprocess related emissions. The assumption made concerning the fuel mix was that therest of the energy (except for the energy from non-replaceable fuels) at the refinery wasproduced by burning refinery gas. Refinery gas has an emission factor of 66.73 tCO2/GJ and a fuel thermal value of 48.15 TJ/ k tonnes8. According to the ReferenceDocument on Best Available Techniques for Mineral Oil Refineries (EuropeanCommission, 2003), the ratio of gas to liquid refinery fuel depends on many factors.The ratio can vary from 80/20 for a stand-alone, moderate complex refinery to 40/60 fora highly complex refinery, which also serves as a chemical complex. It is also said thatthe ratio can be increased when energy conservation measures are applied and the gasavailability becomes sufficient for the energy supply for the refinery. Hence, it can besaid that the assumption that the only fuel used except for non-replaceable fuels isrefinery gas could be true for a highly energy efficient refinery.

8 Source: Revised 1996 IPCC Guidelines for National Greenhouse Gas Inventories; Reference Manual.This is also the source for emission factors and thermal values used by the European Commission(2004/156/EC).


26

4.3.2 Process vents – process-related CO2 emissions

These emissions include process vents and non-replaceable fuels.

4.3.2.1 Non-replaceable fuels

Processes in which there are non-replaceable fuels were considered separately. Theseprocesses include:

• regeneration of catalytic cracker catalyst and other catalysts

• cokers

Regeneration of the catalyst in catalytic cracker and other processesThe regeneration of the catalyst in the catalytic cracker results in a significant amount ofCO2 emissions. The operators were asked for the amount of coke burnt annually. Thereference CO2 emissions from the coke burnt were determined by the conservativeassumption that carbon fraction in the coke burnt was 1.0, and that all of the carbon wasoxidised to CO2. (Most likely the actual carbon fraction is somewhere between 0.9-1.0).Equation 3.2 describes how the CO2 emissions were calculated. Petroleum coke has athermal value of 28.05 GJ/ton (Feldhausen et al, 2004)9. The CO2 emissions from theregeneration of other catalysts were considered negligible in comparison to othersources included and were therefore not considered when calculating the referenceemissions.

CokingOnly fluid cokers and flexi-cokers may have CO2 emissions resulting from the cokeburner. The same equation and calculation methodology can be used as for the catalyticcracking units. It could be assumed for cokers that all of the carbon in the coke isoxidised to CO2 and that the carbon fraction is 1.0. Only two refineries in this study hadcokers other than delayed cokers - one had a fluid coker and another had a flexi-coker.Neither of these two refineries answered the questionnaire, and hence the tier 2methodology was used for estimating the reference CO2 emissions.

4.3.2.2 Process vents

The only process vent considered in this study that is not related to burning of fossilfuels for energy purposes and that can not be considered a non-replaceable fuel ishydrogen production.

9 This is the Swedish specific value and there exist a number of different values for different countries.

The Swedish value has been used for all refineries.


27

Hydrogen production plant.The production of hydrogen results in a significant vent of CO2. A simple approach ofan emission factor of 13.41 tonnes CO2 per million standard cubic feet of H2 producedwas used. This emission factor is given in the API Compendium 2004 and is based on astoichiometric conversion for a feed gas with an average natural gas composition. Theamount of produced hydrogen was asked for in the questionnaire. According to theCommission, the CO2 vent from the hydrogen production plant should be determined byusing data on the composition and amount of feed to the process. The reason for notusing that approach was the lack of data. Data on production was more easilyaccessible.

4.3.3 Actual CO2 emissions

The actual amount of CO2 emissions was given directly by the operator in thequestionnaire.

4.4 Tier 2. Calculation of the CO2 intensity – O&GJ /Solomon

In those cases we did not get a positive response from the refineries (they could notparticipate in the survey by filling out the questionnaires), we used other sources andcalculation methodologies and data sources in order to determine the CO2 intensity. Alldata sources used for this tier are valid for 2001, except for the Solomon study, which isonly performed every second year and hence the 2002 study was used.

Used documents/studies are, amongst others:

• The Solomon study (Solomon Associates 2002), which consists of a fewdocuments:

The process unit energy standard document describing all processes in therefinery and the corresponding process unit energy standard [KBtu/bbl feed](thousands of British Thermal units per barrel feed).

The process statistics document, with statistics for most of the processes (mostcommon) including:

♦ number of installations included in the study,

♦ UUOT (unit utilisation outside turnarounds) in different regions or fordifferent sized installations,

♦ other parameters important for the energy consumption and CO2

emissions.

• The Worldwide Refining survey - every year The Oil and Gas Journal (O&GJ)completes a document called the Worldwide Refining Survey listing all oil


28

refineries in the world. For each refinery, charge/production capacities for the mainprocesses are given. The information from the Worldwide Refining Survey has beenused in order to determine the reference energy demand of the refineries. Each ofthe processes given in the O&GJ (2001) was identified as one of the processes givenin the Solomon e-factor document.

4.4.1 Energy related emissions – combustion of fossil fuels

The reference energy consumption for tier 2 was determined by using the Solomonprocess unit energy standards and production data as described by Equation 3.4. In thisstudy, the load/production (Q) of each process was determined by the charge10 capacityvalues given in the O&GJ 2001(see below) and the Unit Utilisation Outside Turnaround(UUOT) factors given in the Solomon study. Equation 4.4 determined the referenceenergy consumption for each refinery

Equation 4.4 Reference energy consumption for refinery as determined in this study

im

1i UUOTicapacity chargeindardenergy sta unit process SolomonnconsumptioenergyReference ∑ = ⋅⋅=

where, Solomon process unit energy standard [Btu/bbl]Reference energy consumption [Btu/yr]i indicates the processm is the total number of process at the refineryCharge capacity is the size of the process unit [bbl/calendar day]11

UUOT = unit utilisation outside turnarounds, implying Equation 4.5

Equation 4.5. Determination of actual utilisation/load

[ ]bbl/yrloadactualUUOTcapacitycharge =⋅

When determining the reference CO2 emissions according to Equation 4.3 the fuel mixused was process-specific fuels (i.e. coke) and refinery fuel gas just as in the tier 1methodology.

10 Note that for some processes the unit of the process unit energy standard is Btu/product in tonnes. Theproduction capacity should therefore be used instead of charge capacity when determining the energydemand. In order to simplify the reasoning, we used charge capacity although it was production capacityfor some of the processes.11 In O&GJ, the capacity is defined as the maximum number of barrels of input that can be processedduring a 24-hour period, after making allowances for the following: a) types and grades of inputs to beprocessed, b) types and grades of products to be manufactured, c) Environmental constraints associatedwith refinery operations, and d) Scheduled downtime such as mechanical problems, repairs andslowdowns.


29

When determining the reference CO2 emissions from process related fuels and vents,the same methodology was used as in tier 1. The only difference was the estimation ofunit utilisation, which was made by using statistics from the Solomon process statisticsdocument. The UUOT value used was the one for the closest geographical region.

Equation 3.2 was used for calculating the reference CO2 emission from a catalyticcracker. In the calculations here, the coke burn rate was determined by the followingequation (see also the example given in section 3.2.2.5):Equation 4.6. Determination of the coke burn rate in catalytic crackers

ProductionCokeDensityGasFeedUUOTCapacityCrackerCatalyticrateburnCoke ⋅⋅⋅=

where, Catalytic cracker capacity [bbl/yr.] = the value is given in the WorldwideRefining Survey (O&GJ, 2001)

UUOT = the unit utilisation outside turnarounds, given in the Solomon study12.

Feed gas density [kg/m3] = given in the Solomon study12.

Coke production [% wt of fresh feed] = given in Solomon study12.

The carbon fraction of the coke burned and the fraction oxidised was estimated in thesame way as for tier 1. Equation 3.2 was used for calculating the CO2 emissions.

Other processes in a refinery might also have a catalyst that will have to be regenerated.As mentioned in section 4.3, the regeneration of coke in processes other than thecatalytic cracker have been considered negligible in comparison.

Due to lack of process data for cokers, the emissions could not be determined. Only twoof the 89 refineries in this study had cokers other than delayed cokers. The referenceCO2 emissions for those refineries are therefore underestimates.

4.4.2 Process vents – process-specific CO2 emissions

When calculating the reference emissions from the hydrogen production the samecalculation methodology was used as in tier 1, except for the estimation of the unitutilisation, which in tier 2 was estimated by the statistics on process operation in theSolomon study.

4.4.3 Actual emissions

The actual CO2 emissions from the refineries were in most cases taken from the EPERdatabase (http://eper.eea.eu.int/eper/). In some cases, where EPER data was not

12 The average value for Western Europe or Central & South Europe was used, depending on the locationof the refinery.


30

available, data was taken from the national allocation plan of the Member State inquestion.

4.5 Tier 3. Calculation of the CO2 intensity – other data fromproducers

Since the uncertainty of the CO2 intensity based on tier 2 is relatively large (see Section6 below) and only a few of the refineries answered the questionnaire, a thirdmethodology for determining the CO2 intensity was established. The tier 3 methodologyis a mixture of tier 1 and 2. Some of the refinery-specific data used was gathered frominformation on refinery performance and emissions provided by the producers in publicdocuments such as environmental reports. Since this data did not exactly correspond tothe data given by the questionnaires, some extra calculations were necessary in order toget the wanted variables. The data used in this tier are valid for 2001.

5 Data collectionIn order to collect specific data from the refineries, a questionnaire was sent out to 97refineries in the EU15 Member States13 and Norway. Subsequently, bitumen refinerieswere excluded, mostly because the production at those refineries differs from the fuelrefineries (many of them are not included in the Solomon study). The total number ofrefineries included in this study was thus 89. The questionnaires were specific for eachrefinery. 56 of the refineries responded to the circular. 90% of those answered that theycould not participate by filling out the questionnaire. Only five refineries answered thequestionnaire - three Swedish, one Finnish (only part of the questionnaire was filled out)and one German. The reason for not being able to participate differed among thecompanies. In some countries, the process of preparing for the national allocation planwas at a very early stage where the national government not yet had asked theinstallations for emission data. It was then considered a sensitive issue to answer theIVL questionnaire. In some cases, the workload at the refineries was very high and thequestionnaire was not given high priority. However, the majority of refineries/-companies answered that CONCAWE (the European Oil companies’ association forenvironment, health and safety in refining and distribution) ought to be involved.Unfortunately, CONCAWE was not able to join the project under the planned timeframes.

13 Luxembourg has no mineral oil refineries.


31

5.1 Tier 1

The refineries that answered the questionnaire gave the following information used forthe calculations of the CO2 intensity (note that the data asked for in the questionnaireswas valid for 2002):

- Use of fuels and fuel thermal values- Capacity and actual production for the processes given in the O&GJ- Solomon EII- Total amount of CO2 emissions- Amount of electricity purchased- Any extraordinary coincidences occurring at the refinery during the year in question

(2002) that could have altered the amount of emissions significantly

Data of the composite refinery configuration factor and crude oil density was takenfrom the Solomon study. The regional average values were used. The amount of steamand heat purchased or exported was not asked for, but has been extracted from othersources.

5.2 Tier 2

Since only five completed questionnaires were returned, other data sources had to beused. The Oil and Gas Journal annually completes a worldwide survey of refinerycapacity (Stell, 2001). In that study, the charge/production for each of the majorprocesses at each refinery is given. Stell (2001) was used as the source for whichprocesses the different refineries have and what capacity they have. The Solomon studyof energy efficiency at refineries includes almost 200 refineries (85 in Europe) as wellas a lot of process-specific statistics. One parameter used in this study is the unitutilisation outside turnarounds (UUOT), which is given for each process in the Solomonstudy. The value was used together with the charge/production capacity in order toestimate the actual production/load for each of the processes.

In the tier 2 calculations, the Solomon process unit energy standards were used in orderto estimate the reference energy for each process at the refinery. Each of the processesin the O&GJ was connected to a specific process unit energy standard as defined bySolomon Associates. The total amount of reference energy at the refinery was used as abasis for calculating the CO2 emissions. For many of the processes, the data needed forcalculating the total energy needed from the process unit energy standard was the loador the amount produced. However, in some cases this had to be complemented withother statistics of the process, such as the density of input product or coke yield. Allsuch data was taken from the Solomon study.


32

The data on total CO2 emissions during 2001 was taken from the EPER database. Inorder to reduce the uncertainty of that data, some of it was verified by comparing to datagiven in the national allocation plans of the countries.

5.3 Tier 3

In some cases, refinery-specific data was collected from environmental or other reports,or from individual refinery’s web sites.

6 UncertaintiesIn this section, the uncertainties associated with the different calculation methodologiesand tiers are estimated. There are a number of sources to uncertainty in the finaldetermined CO2 intensity index of the individual refineries. The uncertainty of tiers 2and 3 are far greater than the uncertainty of tier 1.

A general uncertainty is the electricity export that many refineries have but for whichwe were unable to obtain data. If a refinery exports electricity, that should be consideredas an extra source of fuel. It was also considered that the heat that many refineriesdeliver does not need extra fuel, but is rather a way of using excess heat. If somerefineries indeed use extra fuel in order to deliver heat, which would cause anunderestimate of the reference CO2 emissions.

It should be noted that the IVL CO2 intensity indices calculated according to tier 1 arevalid for the year 2002 whereas the indices calculated according to tier 2 and 3 are validfor the year 2001.

6.1 Tier 1

The uncertainties associated with this calculation methodology are mainly connected tothe uncertainties in the values given by the operators. The values of composite refineryconfigurations and electricity export also affect the uncertainty. Obtaining data directlyfrom the operators and not using average values could have reduced the uncertaintycaused by those factors. The effect of using the Solomon average value of crude densityis probably negligible.

6.2 Tier 2

There are a lot of assumptions made within this calculation methodology and theuncertainty is much greater than for tier 1. In the tier 2 methodology, where most of thedata was taken from the Worldwide Refining Survey 2001 and the Solomon study, theuncertainty is quite large.


33

The reference CO2 emissions were mainly based on the energy consumption of each ofthe processes at the refinery. Which processes are present at each refinery and of whatsize they are was found in Stell (2001). We learnt by the few answers that we receivedfrom the refinery operators that there are discrepancies between the data given in Stell(2001 & 2002) and the actual refinery data.

We used the data on processes and capacity from Stell (2001) together with data onprocess energy standards and other process-specific data from the Solomon study todetermine the reference energy consumption. A source of uncertainty in thesecalculations was the difference in resolution between the two sources. The WorldwideRefining Survey by Stell (2001) is not as detailed as the Solomon study (there are moredifferent processes) and it was not always obvious which process in Stell (2001)corresponded to which process in the Solomon study. This is considered as one of themain sources of uncertainty in the tier 2 calculations.

It should also be mentioned that the process data and the unit utilisation (utilisation ofcapacity) were based on average values for the refineries in a certain region and not onrefinery-specific data, which further adds to the uncertainty.

Another important data source is the EPER database, from which actual emissions forthe refineries were determined. However, not all refineries could be found in EPER andnot all of them were associated with emissions over the threshold value set in EPER(100 000 t CO2/yr) even though the size of the refinery indicated that the emissions byall means should be greater than the threshold. In order to eliminate large errors in theactual emission data, the EPER data was compared to the historic emissions or theallocation data given in the national allocation plans. Not all countries have includedsuch information in their national allocation plans, but in the cases where it waspossible, this comparison was done. In those cases where it was obvious that the EPERvalue was incorrect, the CO2 emissions were taken instead from the annual report of therefinery.

In the tier 2 calculations, no consideration was given to export of electricity, since suchdata was lacking. For individual refineries, this will strongly affect the CO2 intensityindex (which then falsely might be higher than for refineries with no “excess” electricityproduction).

The CO2 intensity index of some of the refineries determined using the tier 2methodology ended up with values exceeding 250-300. Some of these values areprobably too high due to the uncertainties described earlier in this section and weretherefore excluded when determining the country or regional average.


34

6.3 Tier 3

The tier 3 calculation methodology was only used in a few cases, mostly because of thetime-demanding data collection. The purpose of using this methodology was to reducethe uncertainty compared to the tier 2 methodology. Most of the data used was found onthe web, in environmental reports, or in other descriptions of the refinery operation.

7 ResultsThe objective of this study was partly to assess data on emissions, allocation andproduction for the mineral oil refining industry in Europe. Data on CO2 emissions wasassessed for most refineries included in the study. Some Member States have givenindividual allocation data (as of 2 September 2004) for the refineries and these figurescan be found in Appendix 1 to this report. The IVL CO2 intensity index of the refinerieswas determined, but the values for individual refineries are very uncertain so the sectionbelow presents average values for the different regions within the studied area. Data onproduction, such as the utilisation of the different process units at the refineries, wasrequested in the questionnaires. However, since only a few refineries (five of 97)responded, these data were not compiled and will not be presented here. The fact thatonly a few of the refineries answered the questionnaire resulted in the use of the tier 2calculation methodology of the CO2 intensity index for the majority of refineries. Ifnothing is mentioned, the refineries did not fill out the questionnaire and tier 2calculation methodology was used. The values of the Solomon EII given for the regionsare all valid for 2001. The IVL CO2 intensity indices calculated according to tier 1 arevalid for 2002 and the IVL CO2 intensity indices calculated according to tier 2 and 3 arevalid for 2001 (if not otherwise noted).

7.1 The CO2 intensity for refineries by country

In this section the IVL CO2 intensity indices for the EU15 and Norwegian refineries arepresented. The presentation is made by country and the tier used for the calculations isnoted.

7.1.1 Austria

There is one refinery in Austria – the Schwechat refinery owned by OMV AG. There isan extensive report written by Ecker and Winter (2000) in which the historic emissionsand fuel usage of the Schwechat refinery is described. This report was used to determinethe CO2 intensity index according to a combination of tier 2 and 3. The Schwechatrefinery delivers heat to both Vienna airport and households in Vienna. The informationin Ecker and Winter (2000) also reveals that there are two large power plants producingelectricity both for the need of the refinery and for sale to external users. In the


35

calculations, the amount of delivered electricity was considered.Since there is only one refinery in Austria and individual refinery values are not given,the average value for Austrian and German refineries were combined and the combinedaverage was determined to be 121. In Figure 7.1 the European Inland bar represents theAustrian and German refineries and the Solomon EII for that region is 80.

7.1.2 Belgium

There are five refineries in Belgium, as listed in Table 7.1. Of these five refineries, onlyfour were included in this study. The Nynas Petroleum refinery in Antwerp was notincluded as it is a bitumen-producing refinery with a production quite different fromfuel refineries.

Table 7.1. The Belgian refineries.

Refinery Owner

Antwerp Nynas Petroleum ABAntwerp Belgian Refining CorporationAntwerp ExxonMobilFina Raffinaderij Antwerp TFEUniversal Petroplus

According to the Solomon study, the EII index for the seven refineries included in theBeNeLux countries range between 66-80, with an average of 75. A weighted average(weighted by CDU, Crude Distillation Unit,-capacity) results in the IVL CO2 intensityindex for Belgian fuel refineries being 88. The average CO2 intensity index for theBeNeLux region was determined to be 97 (the Dutch Pernis refinery was excluded14

from the average. If included, the average would have been 119).

7.1.3 Denmark

There are two refineries in Denmark - Fredericia (Shell) and Kalundborg (Statoil), aslisted in Table 7.2. None of the operators filled out the questionnaire, but data requiredfor tier 3 calculations were available on the operators’ websites and that methodologywas used when determining the CO2 intensity index. The electricity export of theFredericia refinery was considered in the calculations.

14 The reason for excluding this refinery was the very high value of the CO2 intensity index received. Thereason for the high value might be that there are processes at this refinery that have not been consideredproperly or that some of the input data for the calculations were inaccurate.


36

Table 7.2. The Danish refineries

Refinery Owner

Fredericia ShellKalundborg Statoil

The Danish refineries were included in the Scandinavian average value of the IVL CO2

intensity index, which was calculated to be 96. The weighted Solomon EII index for theScandinavian refineries is 72.

7.1.4 Finland

In Finland there are two refineries – the Porvoo and Naantali refineries, both owned byFortum Gas Oy, as listed in Table 7.3. The Porvoo refinery partly answered thequestionnaire and some additional information was found on the operator’s website sothat a combination of tier 1/tier 3 methodology could be used when determining the IVLCO2 intensity index. Data required for tier 3 calculations of the Naantali refinery wasfound on the company’s website. In the 2002 calculations for Porvoo, the fact that therefinery sold electricity to external users was considered. The reported 2002 CO2

emissions for the Porvoo refinery were much higher (156 %) than in 2001. Data for2002 was used (i.e., data on actual emissions and on refinery capacity).

Table 7.3The Finnish refineries.

Refinery Owner

Naantali FortumPorvoo Fortum

A weighted average of the IVL CO2 intensity index for the Scandinavian refineries wasdetermined to be 96. The average Solomon EII for Scandinavia refineries is 72.

7.1.5 France

There are 13 refineries in France, as listed in Table 7.4. The ExxonMobil refinery inDunkirk was not included in the survey as that refinery does not have a CDU processand mainly produces lubricant oils, which means that the production is not similar to theaverage of other fuel refineries. The fact that some of the larger refineries lack emissiondata of CO2 in EPER and that some of the calculated CO2 intensity indices were veryhigh indicates that the validity of the EPER data must be questioned. In some cases theemission data given in the French national allocation plan (NAP)was used. Theemission data used is the allocation base, i.e. the average emissions during 1998-2001.According to the list given in the French NAP, there are 14 refineries. However, one ofthem, the SARA Refinery in Le Lamentin, is not included in the O&GJ survey and, due


37

to lack of data, was not included in this study. According to the French nationalallocation plan, it is a small refinery.

Table 7.4 The French refineries

Refinery Owner

Lavera BP PLC (no value of CO2 emissions in EPER, NAP value used)Port Jerome ExxonMobilFosSurMer ExxonMobil (no value of CO2 emissions in EPER, NAP value used)Dunkirk ExxonMobil (not included in this survey)Petite Couronne ShellReichstett ShellBerre L’Etainge Shell (no value of CO2 emissions in EPER, NAP value used)Donges TotalDunkerque/Loon Plage TotalChateauneuf les Martigues Total (no value of CO2 emissions in EPER, NAP value used)Feyzin TotalGrandpuits TotalGonfreville l’Orcher/Harfleur Total

The average Solomon EII for French refineries was 86 in 2001 (min = 77, max = 96),which is above the Western European average (81). The weighted average of the IVLCO2 intensity index for the French refineries was calculated to be 133.

7.1.6 Germany

There are 17 refineries in Germany, as listed in Table 8.5. Of these, 16 included in thisstudy; only the Addinol refinery in Krumpa was not included as there is no CDU andthe refinery mainly produces lubricants. Only one of the refineries filled out thequestionnaire, while most of the others had different reasons for not participating. Theweighted average of the IVL CO2 intensity index for those German refineries where avalue could be calculated and the Austrian refinery was 121. If outliers were excluded,the value would have been 112. The average Solomon EII for the refineries in theEuropean inland is 80.Data on CO2 emissions could not be found for all refineries in the EPER. It is notprobable (due to the size of the refineries) that these refineries have CO2 emissionslower than the threshold value unless they were closed down during a longer timeperiod of 2001. A reason for not having a value of CO2 emissions in EPER might be dueto the division of utilities as separate units. In the list of installations attached to theGerman national allocation plan, many of the refineries are divided into more than oneinstallation. In the German NAP there is so far (2 Sept. 2004) no list with allocation toinstallations or historic emissions. Another reason for unreasonably high or low valuesin the German refineries could be electricity production and gasification projects. Thereis at least one known gasification project among the German refineries.


38

Table 7.5. The German refineries.

Refinery Owner

Vohburg/Ingolstadt/Neustadt BayernoilHamburg BP (no value of CO2 emissions in EPER)Heide/Graasbrook Dea (no value of CO2 emissions in EPER)Wesseling DeaGodorf ShellHarburg ShellIngolstadt ExxonMobil (no value of CO2 emissions in EPER)Salzbergen PharmazeutischeHolborn Holborn Europa RaffinaderiKarlsruhe OberrheinLeuna Mitteldeutsche ErdolBurghausen OMV Mineralölraffinerie Werk BurghausenSchwedt PCK Raffinerie GmbHGelsenkrichen Ruhr Oel GmbHLingen Deutsche BP AktiengesellschaftWillhelmshafener Willhelmshavener

7.1.7 Greece

There are four refineries in Greece, as listed in Table 7.6. The weighted average of the IVLCO2 intensity index of the Greek refineries was calculated to be 146. If the Elefsis refinerywas excluded from the average, the value was reduced to 118. The Elefsis refinery is arefinery of simple structure with only a few processes. For refineries like Elefsis with onlya few processes (no VDU, Vacuum Distillation Unit), it seems to be more important thatthe energy usage of utilities losses and off-sites are excluded in the index. The reasonmight be that they have fewer possibilities to “recycle” heat and efficiently save energy.The average Solomon EII for South & Central European refineries is 105. In the group ofSouth & Central European refineries used by Solomon, there are 13 refineries included.Exactly which countries these refineries are situated in is not explicitly given in theSolomon study. Since all other refineries15 were included in other Solomon-definedgroups, the Greek refineries alone constitute this group in our study.

15 The Solomon group ‘European Inland’ also includes 13 refineries. In our study, we have included theGerman and Austrian refineries, which might not be the same refineries as in the Solomon study. Some ofthese refineries might belong to the group ‘Central & Southern Europe’.


39

Table 7.6 The Greek refineries.

Refinery Owner

Aspropyrgos Hellenic PetroleumThessaloniki Hellenic PetroleumCorinth Motor Oil HellasElefsis Hellenic Petroleum

7.1.8 Ireland

In Ireland there is only one refinery - the Whitegate refinery owned by ConocoPhilliphs. The average IVL CO2 intensity index for British and Irish refineries was 127.The average Solomon EII for British and Irish refineries is 87.

7.1.9 Italy

There are 17 refineries in Italy, as listed in Table 7.7. The EPER data of CO2 emissionsfrom Italian refineries does not seem to be complete. In some cases where the calculatedIVL CO2 intensity index is very high (or low), there might be special processes orproduction at the individual refinery that need to be considered. The average SolomonEII for Italian refineries is 81. The weighted average of the IVL CO2 intensity index forthe Italian refineries, for which the index was calculated, was 159. If outliers wereexcluded (Gela Ragusa, Falconara and Sarroch) the average IVL CO2 intensity indexwas 135. At Falconara, Priolo Gargallo and Sarroch there are gasification projectspresent. No list of installations with historic emissions is available in the Italian nationalallocation plan (2 Sept. 2004).


40

Table 7.7 The Italian refineries.

Refinery Owner

Milazzo Agip/EniPriolo Siracusa Agip/EniGela Ragusa Agip/EniSannazzaro Agip/EniLivorno Agip/EniTaranto Agip/EniPort Maghera Agip/EniFalconara API SpALa Spezia Arcola Petrolifera (no value of CO2 emissions in EPER)Priolo Gargallo ISAB SpAAugusta ExxonMobilBusalla Iplom SpASarroch SarasCremona TamoilTrecate SarpomMantova Italiana EnergiaRaffineria di Roma TFE

7.1.10 The Netherlands

There are six refineries in the Netherlands, as listed in Table 7.8. Only five of therefineries were included in this survey. The Smid & Hollander refinery was omitted as itis a bitumen refinery and not a fuel refinery. A weighted average of the Dutch andBelgian (BeNeLux) IVL CO2 intensity index was determined to be 97 (when outlierswere excluded). The IVL CO2 intensity index for the Dutch refineries was only 105(excluding outliers, see footnote 14). According to the Solomon study, the EII index forthe seven refineries included in the BeNeLux countries range between 66-80 with anaverage of 75.

Table 7.8. The Dutch refineries.

Refinery Owner

Rotterdam ExxonMobilRotterdam Q8Europoort NerefcoPernis ShellAmsterdam Smid & Hollander (not included, bitumen refinery)Vlissingen Total


41

7.1.11 Norway

Norway was included in this survey although it does not belong to the EU15 MemberStates. The reasons for the inclusion were because Norway has shown interest inparticipating in the EU ETS from the beginning in 2005 and because they arecompetitors to most of the EU15 refineries. There are two fuel refineries in Norway, aslisted in Table 7.9. The weighted average IVL CO2 intensity index of the Scandinavianrefineries was calculated to be 96 and the Solomon EII for the region is 72.

Table 7.9The Norwegian refineries.

Refinery Owner

Slagen ExxonMobilMongstad Statoil

7.1.12 Portugal

There are two refineries in Portugal - the Porto and Sines refineries both owned by GalpEnergia, as listed in Table 8.10. The average value of the IVL CO2 intensity index forthe refineries on the Iberian Peninsula was calculated to be 140.

Table 7.10 The Portuguese refineries

Refinery Owner

Porto Galp EnergiaSines Galp Energia

The average Solomon EII index of the 11 refineries included in the Solomon study onthe Iberian Peninsula is 88.

7.1.13 Spain

There are nine refineries in Spain, as listed in Table 7.11. The Spanish refineries have afew processes for which no statistics on UUOT were available in the Solomon study.The average Solomon EII for the refineries on the Iberian Peninsula is 88. A weightedaverage of the IVL CO2 intensity index of the Spanish and Portuguese refineries wascalculated to be 140. Two of the Spanish refineries were excluded when calculating theaverage because the intensity values for those refineries were very high, most likely dueto lack of or erroneous data. Had those two refineries been included, the CO2 intensityfor the Iberian refineries would have been 160. There are gasification projects at two (atleast) of the Spanish refineries (SFA Pacific, 2000).


42

Table 7.11. The Spanish refineries.

Refinery Owner

Castellon de la Plana BPLa Rabida, Huelva CEPSACadiz, Gibraltar CEPSATenerife CEPSACartagena Repsol YPFLa Coruna Repsol YPFPuertollano Repsol YPFSomorrostro Repsol YPFTarragona Repsol YPF

7.1.14 Sweden

In Sweden, there are three refineries that were included in this study. There are also twoadditional refineries that mainly produce bitumen. The included Swedish refineries arelisted in Table 7.12. All three refineries filled out the questionnaire and the IVL CO2

intensity indices were determined according to tier 1and, hence, are valid for 2002.

Table 7.12. The Swedish fuel refineries.

Refinery Owner

Göteborg PreemGöteborg ShellLysekil Scanraff

The weighted average of the IVL CO2 intensity index of the Scandinavian (Sweden,Norway, Finland and Denmark) refineries was 96. The weighted mean of the SolomonEII index for the Scandinavian refineries is 72.

7.1.15 The United Kingdom

There are nine fuel refineries in the United Kingdom, as listed in Table 7.13. There arealso two bitumen refineries in the UK the Eastham and the Dundee refineries. Theywere both excluded in this study since they have a significantly different production.The IVL CO2 intensity index of the UK refineries was calculated using mainly tier 2 butsome extra information on processes was available from the Institute of Petroleum. Theaverage CO2 intensity for the Irish and British refineries was calculated to 128.However, the individual values for some of the refineries seem to be high and might bedue to own production of electricity.


43

Table 7.13. The UK fuel refineries.

Refinery Owner

Fawley ExxonMobilStanlow ShellGrangemouth BPLindsey TotalFinaElfPembrok Chevron TexacoCoryton BPHumber Conoco PhillipsMilfordHaven TotalFinaElfTeesside Petroplus

7.2 Summary of the IVL CO2 intensity index

Figure 7.1 below shows the result of the IVL CO2 intensity index calculations.Countries with only one or a few refineries were grouped (as described in the previoussections) and an average value for each region was calculated.

A correlation between CO2 and energy intensity can be seen. Countries with lowenergy-intensive refineries also have low CO2-intensive refineries (some deviation). Theuncertainty is probably lower for the Scandinavian CO2 intensity index as tier 1 and 3methodologies were applied when determining the indices. All Swedish refineries filledout the questionnaire; the Danish refineries were treated by using tier 3 calculations.Tier 3 methodology was also partly used for the Finnish refineries.


44

CO2 and energy intensity of EU15 and Norwegian fuel refineries

121

97

146

127

140

8086

105

87 8881

135

96

133

75 72

0

20

40

60

80

100

120

140

160

Inland Europe BeNeLux Scandianavia France South &CentralEurope

Ireland &Britain

Iberia Italy

CO2 intensity

Solomon EII

Figure 7.1. The CO2 intensity index as determined by IVL and the energy intensity index asdetermined by Solomon Associates. The denotation of the groups of countries is the oneused by Solomon. The groups made when determining the average IVL CO2 intensityindices are: Inland Europe = Germany and Austria; BeNeLux = Belgium and theNetherlands (there are no refineries in Luxembourg); Scandinavia = Denmark, Finland,Norway and Sweden; South & Central Europe = Greece; Iberia = Spain and Portugal.

7.3 Allocation to the Mineral Oil Refining Industry

The national allocation plans of twelve of the EU25 countries have been analysed inZetterberg et al (2004). In that analysis, special attention was given to the energy andmineral oil refinery sectors. Much of the information given in this section was given byZetterberg et al (2004).

The EU15 allocation plans analysed in Zetterberg et al (2004) were for Austria, Belgian(only draft version of Flandern), Denmark, Germany, Finland, Ireland, Luxembourg, theNetherlands, Portugal (draft version), Sweden and the United Kingdom. Figure 7.2 onallocation vs. historic emissions and projected emissions is a modification of Figure 4.2in Zetterberg et al (2004), where additional information given by the countries to theCommission has been added. However, Germany has not given current emissions of therefining sector and Greece has still not published a national allocation plan (2 Sept.2004). Belgium, France, Italy and Spain submitted their allocation plans very late (theywere not included in the analysis made by Zetterberg et al 2004) and therefore onlydata, and no allocation methodology, were extracted from those allocation plans.


45

Allocation vs Current and Projected Emissions

1,071,02 1,05 1,09

1,011,07 1,05 1,091,10 1,11

1,11 1,15

0,96 1,011,070,92

0,0

0,2

0,4

0,6

0,8

1,0

1,2

1,4

1,6

Austria

Belgium

Denmark

Finlan

d

France

Irelan

dIta

ly

Netherl

ands

Portug

alSpa

in

Sweden

United

King

dom

Allocation / Emission (latest year available)

Allocation/Projected Emissions

Figure 7.2. The allocation vs. current and projected emissions to the mineral oil refining sectors indifferent Member States. Unfortunately, not all Member States gave data on current and/orprojected emissions for the mineral oil refining sector.

Generally, in most countries the refinery sector has been allocated more allowances thancorresponding current emissions. However, the allocation includes both allowances toexisting installations, new entrants and to expansions. Some large newentrants/expansions are known, such as the Gasification Project at Scanraff refinery inSweden and the upgrade at the Porvoo refinery in Finland. It was only in the Swedishcase that the amount of allowances allocated to the new installation/expansion could beidentified and excluded from the amount of allowances allocated in Figure 7.2. In theFinnish case, the allowances allocated for the expansion/upgrade could not bedistinguished in the allocated amount and are therefore included in Figure 7.2. Theremight also be other known new installations or expansions in other countries that nothave been excluded, making the allocation appear to be more beneficial for the refinerysector in those countries than it is in reality. It should also be noted that what isconsidered as projected emissions differs between countries. Only four countries gavevalues for projected emissions for the mineral oil refining sector (2 Sept. 2004). It hasnot been decided formally whether or not Norway will link up with the EU ETS (2 Sept.2004).

In Figure 7.3, both the quotas for allocation/historic emissions and the IVL CO2

intensity indices for the refining sector in the countries included in this study areplotted. Note that not all of the original EU15 Member States are included in Figure 7.3since not all countries have published historic emissions and/or allocation to the refiningsector specifically (2 Sept. 2004). Germany, Greece and Norway (which is not aMember State) are not included due to these reasons. The Member States have been


46

plotted in order of increasing IVL CO2 intensity index. A higher value of the IVL CO2

intensity index means higher emissions of CO2 per produced unit compared to a lowervalue of the IVL CO2 intensity index.

Allocation quota vs CO2 intensity index

101102

105

107 107109 109

110 111 111

115

120

90

95

100

105

110

115

120

125

Portug

al

Belgium

France

Austria

Spain UK

Irelan

dIta

ly

Netherl

ands

Denmark

Finlan

d

Sweden

Allo

catio

n/hi

stor

ic e

mis

sion

0

20

40

60

80

100

120

140

160

CO

2 int

ensi

ty in

dex

Allocation/historic emissions (lates year available)

CO2 intensity

Figure 7.3. The average IVL CO2 intensity index of the refining sectors in different Member States and theratio between allocation and historic emissions. Note that the regional values of the IVL CO2

intensity indices have been used for countries with less than four refineries. That means that thevalues of the IVL CO2 intensity indices for Denmark, Sweden and Finland are represented bythe Scandinavian average. Correspondingly, the Austrian value is represented by the EuropeanInland value, the Spanish and the Portuguese values are represented by the Iberian value andthe Irish and the British values are represented by the Ireland & Britain average value. Notethat in this figure, the ratio of allocation and historic emissions for Sweden also includes a newlarge Gasification project at the Scanraff refinery.

If the CO2 intensity/efficiency had been rewarded on a European level when allocatingto the mineral oil refining sector, relatively less allowances would have been allocatedto CO2-intense installations. Or in terms of Member States, those with a CO2-intenserefining sector would have allocated relatively less allowances to that sector thancountries with less CO2-intensive mineral oil refining sectors would have. That meansthat the bars and the curve in Figure 7.3 should be negatively correlated, i.e., as the barsget higher, the curve should decline. As can be seen, this is the case for the refiningsectors in some countries, but not all.

Figure 7.4 is a slight modification of Figure 7.3, where the allocation to the largeexpansion at the Scanraff refinery in Sweden was excluded (since there are no historicemissions from this installation that will be commissioned in 2006). Even though it


47

seems as if the country with the lowest allocation compared to current emissions is alsoamongst the countries with the most CO2-intensive refinery sector, i.e. Portugal, thereare also countries with a relatively CO2-efficient refining sector that do not get muchhigher allocation (relatively), such as Belgium (and Sweden according to Figure 7.4).There are also other countries with relatively high CO2 intensity, such as UnitedKingdom, Ireland, Italy and Spain that have allocated relatively high amounts ofallowances to their refining industry. It can also be seen that the Scandinavian countries(except Sweden according to Figure 7.4) and the Netherlands, who have relatively CO2-efficient mineral oil refining sectors, have allocated relatively high amounts to theirsectors. Further, France, with a relatively low CO2 efficiency has allocated relativelyfew allowances to its mineral oil refining sector.

Criterion 5 of Annex III to the Directive concerns the competition between companiesand sectors. In order to avoid differences in allocation methodology between companieswithin the same sector, the Commission could have compared the allocation to the samesector between Member States. However, the national allocation plans have beenindividually scrutinised by the Commission and sectors have been treated differently indifferent Member States.

The allocation in most cases includes allowances to new installations. In other words, ifthe sector is expanding, the allocation to the sector should also be higher than thehistoric emissions. For example, in the case of the Swedish and the Finnish refinerysector, we know that the allocation includes a considerable amount for new installationsor expansions. If the allowances allocated to new installations are excluded, the pictureof the allocation vs. historic emissions would look different. This can be seen bycomparing Figure 7.4, where the allocation to the large expansion at the Scanraffrefinery in Sweden was excluded, and Figure 7.3. The exclusion of the allocated amountto the large expansion in Sweden makes the figure look quite different and, mostprobably, the Finnish bar would also be diminished radically by excluding the allocationto the upgraded part of the Porvoo refinery. There might also be correspondingadjustments to be made for other countries that are not known to us. There have beendifficulties in distinguishing between allocation to present installations and newinstallations or expansions. The Commission asked many countries to makeclarifications and amendments to their allocation plans. It is therefore possible that theCommission had access to more information when scrutinising the allocation plans thanwe have been able to obtain.


48

Allocation quota vs CO2 intensity index

101102

105 105

107 107109 109

110 111 111

115

90

95

100

105

110

115

120

Portug

al

Belgium

Sweden

France

Austria

Spain UK

Irelan

dIta

ly

Netherl

ands

Denmark

Finlan

d

Allo

catio

n/hi

stor

ic e

mis

sion

0

20

40

60

80

100

120

140

160

CO

2 int

ensi

ty in

dex

Allocation/historic emissions (lates year available)

CO2 intensity

Figure 7.4. Modification of Figure 7.3. The new entrants have been excluded in the Swedish amount ofallocated allowances.


49

8 Discussion

8.1 The differences in CO2 intensity between refineries

There are a few important factors that determine the difference in IVL CO2 intensityindex between refineries. One of them is the difference in fuel mix. In Figure 8.1, theweighted emission factors show differences in CO2 intensity of the fuel mix used atrefineries in different countries. Note that purchased electricity, steam or heat is notincluded. Even if the refining sector in a country is relatively energy efficient, the use ofmore CO2-intense fuels will make them less CO2 efficient.

Weighted emission factor

76,4 76,4

72,5

75,3

75,9

73,8

72,3

74,5

70

71

72

73

74

75

76

77

Central &SouthernEurope

Britain andIreland

France Inland Europe BeNeLux Scandinavia Italy Iberia

g C

O2/M

J

Figure 8.1. The weighted emission factor is a based on the actual fuel mix at the refineries in differentregions. Data source: Solomon Associates (2002).

The total amount of used energy and the amount of purchased energy (steam andelectricity) also affect the CO2 intensity index, as described earlier. Figure 8.2 shows theenergy consumption per utilised equivalent distillation capacity (EDC), a measure oftotal refinery utilisation, at the refineries in different regions in Europe.

The amount of purchased electricity affects the CO2 intensity since purchased electricityis not associated with any CO2 emissions. Figure 8.3 shows the amount of purchasedelectricity and produced electricity at refineries in different regions in Europe.


50

Total Energy Consumption, GJ/yr per Utilized EDC

16,8 17,3

12,6

14,0

19,1

14,1 14,1

16,1

0,0

5,0

10,0

15,0

20,0

25,0


Britain andIreland


GJ/

yr p

er U

tiliz

ed E

DC

Figure 8.2 Total energy consumption [GJ] per utilised EDC (equivalent distillation capacity) atrefineries in different regions in Europe. Note that the complexity of the refinery and theproduct slate will affect this factor. Source: Solomon Associates (2002).

Due to data availability, the calculated IVL CO2 intensity indices in this report areaffected by the electricity production at the refineries. In most cases, it was not possibleto consider the extra CO2 emissions needed for the exported electricity due to lack ofdata. If more detailed data had been available and the IVL CO2 intensity index couldhave been determined according to the tier 1 methodology, this problem would havebeen eliminated. However, the amount of exported electricity does not have a greatimpact on the regional level since the refinery sector is not a net seller of electricity inany of the regions. Still, there might be individual refineries producing a lot ofelectricity and for which the impact on the CO2 intensity will be of importance.

Together, the fuel mix, the amount of produced electricity and the amount of energyused per utilised EDC at the refineries in the different countries explain a large portionof the pattern of different IVL CO2 intensity indices of refineries in differentcountries/regions. This is shown in Figure 8.4. However, it does not explain the wholedifference. Also, the product slate of the refineries will be of importance. The compositerefinery configuration (i.e., the complexity of the refinery) is partly a description of thepotential product slate and is a measure that varies significantly between regions. Oneshould also remember that there are large uncertainties in the calculated values of theCO2 intensity for the individual refineries. Since there are Member States with only afew refineries (less than four), even the national average values are sometimes quiteuncertain. In some countries with only a few refineries, a large error in the index willstill make a large impact on the average. On the higher aggregated level, the trends areprobably more certain, which is why we have used the aggregated averages in thecomparison to the allocation.


51

Electricity usage at refineries

0,0

5,0

10,0

15,0

20,0

25,0

AverageEurope


Britain andIreland

France InlandEurope

BeNeLux Scandinavia Italy Iberia

%

Produced electricity of total energy

Purchased electricity of total energy

Figure 8.3. Electricity usage and production at European refineries. Amount of electricity that is used,purchased and produced by the mineral oil refining industry in different European regions.Source: Solomon Associates (2002).

218 215

167

203

156140

191194

0

50

100

150

200

250


Britain andIreland


kg C

O2 /

tonn

e ne

t inp

ut

Figure 8.4. Amount of CO2 potentially emitted per tonne net input. Calculated as (1-ratio of energyusage that is purchased electricity) · weighted emission factor · energy usage per tonne netinput. Source: Solomon Associates (2002).


52

8.2 For what purposes can the IVL CO2 intensity index beused?

The IVL CO2 intensity index has many of the advantages that are also associated withthe Solomon Energy Intensity Index (EII). For instance, it considers the differences incomplexity between refineries. But there are also other advantages that are not includedin the Solomon EII.

• The index is directly related to CO2 emissions (not only energy consumption).

• The index considers process-related emissions (both fuels and process vents) thatcannot be easily substituted.

• The index reveals differences in the use of CO2-intensive fuels. Even though arefinery is energy efficient, more CO2-intensive fuels can be at use, which can beindicated by the index.

• The index can easily be expanded to include greenhouse gases other than carbondioxide, making it a greenhouse gas intensity index. This might be useful when/ifthe EU Emission trading scheme is expanded to include more gases.

However, the uncertainty of the index such as calculated in this study is great and inorder to use the index more widely, the uncertainty will have to be reduced. This can bedone by using refinery-specific data given by, for instance, the refineries themselves.

The principles of the IVL CO2 intensity index could be used as guidance for aninternational benchmark for allocating to mineral oil refineries. Of course, it wouldrequire refinery-specific data, but the majority of the refineries in the EU15 are part ofthe Solomon study and therefore have access to the data used in this study to calculatethe IVL CO2 intensity index. Solomon Associates has also made a study on agreenhouse gas performance index for mineral oil refineries that has a lot in commonwith the IVL CO2 intensity index presented in this study (Solomon Associates, 2003).An international benchmark for the mineral oil refining sector used in the allocationwithin the EU ETS would reward CO2-efficient refineries.

8.3 Expectations and Outcome of the EU ETS nationalallocation plans

One of the main reasons for conducting this study was to see if some refineries weretreated more favourably than others in the allocation process. In particular, we wantedto see if refineries with low energy intensity or low CO2 intensity were treated morefavourably than less efficient refineries.


53

According to the criteria for the national allocation plans, a list of installations should beincluded in the ETS, together with the quantities of allowances intended to be allocatedto each. The plans should also contain the total number of allowances to be distributedwithin the trading sector and the methodology used when determining the number ofallowances to each installation. From this study and the analysis of the nationalallocation plans by Zetterberg et al (2004), it was clear that:

• Few of the national allocation plans were submitted in time to the Commission

• Only a few of the national allocation plans actually included a list of installationsand the number of allowances intended to be allocated to them. Note that some ofthe national allocation plans might have been complemented on request of theCommission. (2 Sept. 2004).

• The description of allocation methodologies used was not easy to penetrate. In manycases, a growth factor was used when determining the amount of allowances to beallocated, but what that growth factor represented exactly was sometimes difficult tosee.

These circumstances are some of the reasons that have made the system less transparentthan we hoped for (and that originally was announced in the Directive) and explain, inpart, why the CO2 intensity index turned out to be quite uncertain at the installationlevel. However, it should be noted that we have not seen the answers to theclarifications required by the Commission before approving the allocation plans.

Based on the analysis of the national allocation plans, it can generally be concluded thatthe mineral oil refining industry has been allocated allowances corresponding to higheremissions than the current emissions. Both new installations and Directive 2003/17/ECconcerning the production of low-sulphur liquid fuels should be considered, as thesewill in fact increase CO2 emissions at refineries (but reduce the emissions forconsumers). Many countries have taken Directive 2003/17/EC into consideration whenallocating to the mineral oil refineries, but not all. In Germany, for example, a refinerywill only be considered in the allocation process if it can prove that the Directive resultsin increased emissions by at least 10%. It should also be noted that some countriesallocate more allowances than the projected emissions.

8.4 Future challenges

The time schedule for implementing the EU ETS has been very tight. However, theCommission has now approved the majority of the national allocation plans and allcountries should have issued the emission allowances to the installations by the end ofFebruary 2005. This means that one very important first step towards a functioningemission trading system has been taken. Verifying the emissions will probably be thenext big challenge for the EU ETS. Each Member State will individually determine how


54

the measurements at the different installations should be performed. There is a guidancedocument from the Commission but it allows for interpretation by the Member States.In the guidance document (2004/156/EC) on monitoring and reporting, it is said thatinstallations with emissions > 500 000 tonnes CO2/a should monitor with an uncertainty< 1.5%. Many of the refineries fall into this category, but it will be a great challenge toput measuring techniques in place in order to fulfil the requirement of such lowmeasurement uncertainty. Improving the measurement uncertainty of the emissions atmany of the emitting installations and thereby declining the overall uncertainty of thetotal emissions is also one of the great advantages of the EU ETS.

Since the allocation to the trading sector in general has been generous (Zetterberg et al2004), the impacts might, at least on the refining sector, not be very great during thefirst period 2005 - 2007. However, the system is now in place and it has great potentialto regulate the emissions from installations included in the future. The importance ofobjectivity and transparency will therefore be even greater in the future. If internationalbenchmarks were used across the EU for sectors or installations when allocating, theparticipants might have found that they were treated more equally. Such a system wouldalso motivate installations to reduce their CO2 intensity if they knew that the nextallocation was going to be based on benchmarks instead of historic emissions.


55

9 ConclusionsIt can be concluded that there are substantial differences in the IVL CO2 intensity indexof the mineral oil refineries in different regions within Europe and that these differenceshave not been considered in the allocation process. However, for some countries there isstill some correlation between allocation and CO2 efficiency.

The uncertainty of the values of the IVL CO2 intensity indices for the individualrefineries is large, especially in those cases where the tier 2 and 3 methodologies havebeen used (97% of the cases). Still, the general trends that could be discovered whencomparing the CO2 intensities of different countries/regions are probably valid.

From the analysis of the national allocation plans it can generally be concluded that themineral oil refining industry has been allocated allowances corresponding to higheremissions than its current emissions. It should be noted that the allocation also includesfuture expansions and new installations, which means that for individual installations,the allocation might not be higher than current emissions.

It can be concluded that a few countries (of those whose national allocation plan havebeen evaluated) have considered energy or CO2 efficiency in the allocation process.Austria, Germany and the Netherlands have mentioned energy efficiency or reductionpotential due to CO2 intensity of fuels used. A few countries also mention early actionalthough most countries have only considered that by using historic emissions as a basisfor allocation. Most countries have considered co-generation of heat and electricity as aclean technology, rewarding better allocation than other production methods. However,no country has explicitly used an efficiency or intensity index for energy or CO2 atrefineries. This is not beneficial for energy and CO2-efficient refineries since they havefewer possibilities to reduce emissions than less efficient refineries do.

Most countries have based the allocation on historic emissions and an estimated growthfactor. Basing allocation on historic emission could be beneficial for refineries that havemade emission-reducing measures in recent years. However, most countries have usedemissions in recent years as allocation basis, which diminishes these benefits.

Only Denmark has explicitly given a value of a benchmark that will be used forallocation to new mineral oil refineries. The use of benchmarks will always bebeneficial for CO2-efficient installations.


56

10 Further research/improvementsIn order to improve the CO2 intensity index, the uncertainty should be reduced. This isprobably best done by a closer co-operation with the refining industry. Additional datathat could be useful are the heat and steam balance (export/import) and the factors takenfrom the Solomon study (composite refinery configuration factor). A well-developedCO2 intensity index could be used as guidance for an international benchmark whenallocating to refineries. This would also help the sector to make the cheapest CO2

emission-reducing measures while not affecting the regional competition.

The CO2 efficiency and environmental advantages of producing energy at refineriescould be investigated further. In so-called gasification projects where IntegratedGasification Combined Cycle technology is used, refineries use low-grade fuels forproducing steam, hydrogen and electric energy with the highest conversion efficiencypossible (European Commission, 2003). These should probably be given more “credits”in the CO2 intensity index. One could also consider heat delivery in a different way,especially if extra fuel is used with the primary objective to increase heat delivery.


57

11 References

11.1 LiteratureAllocatieplan CO2-emissierechten 2005 t/m 2007. Nederlands nationaal toewijzingsplan inzake

de toewijzing van broeikasgasemissierechten aan bedrijven. 16 April 2004.

Danish National Allocation Plan, Ministry of the Environment, March 2004.

Directive 2003/87/EC of the European Parliament and of the Council of 13 October 2003establishing a scheme for greenhouse gas emission allowance trading within theCommunity and amending Council Directive 96/61/EC

Ecker, A., Winter, B., 2000. Stand der Technik bei Raffinerien im Hinblick auf die IPCC-Richtlinie. Umweltbundesamt - Federal Environment Agency Austria. Monographien,Band 119.

European Commission, 2003. Integrated Pollution Prevention and Control IPPC – ReferenceDocument on Best Available Techniques for Mineral Oil and Gas Refineries, February2003. Document available at http://eippcb.jrc.es/pages/FActivities.htm .

European Commission, 2004/156/EC. Establishing guidelines for the monitoring and reportingof greenhouse gas emissions pursuant to Directive 2003/87/EC of the EuropeanParliament and of the Council. Commission Decision of 29 January 2004.

Feldhausen, K., Hammarsköld, G., Mjureke, D., Pettersson, S., Alopeaus Sandberg, T., Staaf,H., Svensson, U., Österberg, K. 2004. Sweden’s National Inventory Report 2004-submitted under the United Nations Convention on Climate Change. SwedishEnvironmental Protection Agency. Stockholm, Sweden. [www.naturvardsverket.se].

Finnish Proposal for National Allocation Plan for Greenhouse Gas Emission Allowances for theyears 2005-2007. Ministry of Trade and Industry, Draft of 30 March 2004.

Furimsky, E., Gasification in Petroleum Refinery of 21st Century. Oil & Gas Science andTechnology – Rev. IFP, Vol. 54 (1999), No 5 pp 597-618.

Integrated Pollution Prevention and Control, IPPC. Reference Document on Best AvailableTechniques for Mineral Oil and Gas Refineries. European Commission, February 2003.

Irelands National Allocation Plan 2005-2007, As notified to the Commission, 31 March 2004.

National Allocation Plan for Austria pursuant to Art. 11 of the EZG, 31 March 2004 withadditions dated 7 April 2004. Federal Ministry of Agriculture, Forestry, Environment andwater Management.

National Allokeringsplan for Danmark, Miljøministeriet, Marts 2004.

Nationaler Allokationsplan für die Bundesrepublik Deutschland 2005-2007, Bundesministeriumfür Umwelt, Naturschutz und Reaktorsicherheit. 31. März 2004, Berlin.


58

National Allocation Plan for the Federal Republic of Germany 2005-2007, Federal Ministry ofthe Environment, Nature Conservation and Nuclear safety. 31 March 2004, Berlin,Translation 7 May 2004.

Nationaler Zuteilungsplan für Österreich gemäss § 11 EZG, 31. März 2004 mit Ergänzungenvom 7. April 2004. Bundesamt für Land-und Fortwirtschaft, Umwelt undWasserwirtschaft.

Plano Nacional de Atribuição de Licenças de Emissão de CO2 (PNALE) 2005-2007. Versäopara Discussão Pública. Ministério da Economia, Ministéro das Cidades Ordenamento doTerritório e Ambiente, 17 de Marçao de 2004.

SFA Pacific, 2000. Gasification - Worldwide Use and Acceptance. Report prepared for USDepartment of Energy, Office of fossil Energy, National Energy Technology Laboratoryand the Gasification Technologies Council. Contract DE-AMO1-98FE65271.

Shires, T., Loughran, C., 2004. Compendium of Greenhouse Gas Emissions EstimationMethodologies for the Oil and Gas Industry. URS Corporation, Texas. Developed for:American Petroleum Institute, Washington, USA.

Solomon Associates 2002. Europe, Africa and the Middle East Fuels Refinery PerformanceAnalysis.

Solomon Associates, 2003. GHG Emission Performance Measurement: The Way Forward. HSBSolomon Associates, LLC.

Stell, Jeannie, Survey Editor. 2001 Worldwide Refining Survey, Oil and Gas Journal, December2001.

Stell, Jeannie, Survey Editor. 2002 Worldwide Refining Survey, Oil and Gas Journal, December2002.

Sveriges nationella allokeringsplan, Regeringskansliet, Näringsdepartementet, Promemoria2004-04-22.

Zetterberg, L., Nilsson, K., Åhman, M., Kumlin, A-S., Birgersdotter, L., 2004. IVL B-1591.Unpublished. Analysis of National Allocation Plans of the EU ETS.

Åhman , M., Zetterberg, L, 2004. Options for Emission Allowance Allocation under the EUEmissions Trading Directive. Accepted Manuscript by Mitigation and AdaptationStrategies for Global Change. Prepublication Date 2004-09-08.


59

11.2 Personal communications

Brinck, Leif, Environmental Manager, Scanraff Refinery 2004.

Magill, Thomas, UOP Limited (Universal Oil Products), 2004

Trout, Bill, Project Manager, Solomon Associates 2004.

Shires, Terri, URS Corporation, Austin 2004.

11.3 Web-sites

http://europa.eu.int/comm/environment/climat/emission_plans.htm

Appendix 1Country Refinery Emissions according to

EPER[ton CO2 in 2001]

Emissions according toNAP [ton CO2/yr]

Allocation[ton CO2/yr]

Comments

Austria Raffinerie Schwechat 2 560 000 2 587 000Belgium Belgian Refining Corporation 462 000Belgium Esso Raffinaderij Antwerpen 1 610 000Belgium Fina Raffinaderij Antwerpen 3 230 000Belgium Nynäs < threshold Not included in surveyBelgium Petroplus Refining Antwerp 191 000Denmark Dansk Shell A/S, Fredericia 263 000 Emissions according to

Environmental Report: 263 000Denmark Statoil A/S, Kalundborg 486 000Finland Fortum Oil and Gas Oy, Porvoo 1 480 000 2 707 877 Emissions according to owner

(2002): 2 312 000 ton CO2Finland Fortum Oil and Gas Oy, Naantali 385 000 352 956 Emissions according to owner

(2002): 355 197 ton CO2

France Raffinerie de Lavera (BP Lavera SNC) < threshold 1 568 498France Shell Couronnaise Raffinage 1 390 000 1 454 715France Shell, CRR Compagnie Rhenane de Raffinage 587 000 604 208France Esso Raffinage FosSurMer < threshold 857 114France ExxonMobil Port Jerome 2 870 000 2 528 449France Raffinage de Provance site de la Mede, Total,

Chateauneuf les Martigues< threshold 1 554 921

France ExxonMobil Sté de la Raffinerie deDunkerque

650 000 265 008

France Total Raffinerie Gonfreville l'Orcher Harfleur 2 380 000 3 239 098France Total Feyzin 1 240 000 1 279 317France Total Fina Elf Dunkuerque, Loon Plage 1 230 000 1 245 772France Total Fina Elf Raffinerie de Donges 1 320 000 1 369 939France Total Fina Elf Raffinerie de Grandpuits 688 000 780 704France Shell Berre L'Etainge 1 603 000 1 304 951

Allowance Allocation and C

O2 intensity of the EU

15 and Norw

egian refineries IVL Report B1610

60

Country Refinery Emissions according toEPER

[ton CO2 in 2001]



Comments

Germany Krumpa, Addinol Not included in surveyGermany BP Hamburg Could not be found in

EPERNo value of CO2 intensity could

be calculatedGermany DEA MineralOel Wesseling 2 150 000Germany Deutsche Shell GmbH Godorf 1 630 000Germany Ruhr Oel GmbH, Gelsenkirchen 3 393 000Germany Bayernoil, Vohburg/Ingolstadt/Neustadt 1 732 000Germany Deutsche Shell GmbH Raffinerie Zentrum

Harburg850 000

Germany Erdölraffinaderie Emsland GmbH & Co KG,Lingen

1 160 000

Germany Esso Deutschland – Raffinerie Ingolstadt < threshold No value of CO2 intensity couldbe calculated

Germany MIRO Mineralölraffinerie Oberrhein 2 628 000Germany Holborn Europa Raffinaderie; Hamburg 491 000Germany MIDER Mitteldeutsche Erdoel Raffinerie,

Leuna, Spregau1 010 000

Germany OMV Werke Burghausen 803 000Germany Raffinerie Schwedt PCK 3 640 000Germany H&R, Schmierstoff Raffinaderie Salzbergen 184 000 Emissions according to owner in

2002: 175 107 ton CO2Germany Shell & Dea Oil GmbH Raffineri Heide/

Graasbrook< threshold No value of CO2 intensity could

be calculatedGermany Willhelmshaven, European Vinyls

Corporation (Deutschland)715 000

Greece Hellenic Petroleum, Asporpyrgos 1 500 000Greece Hellenic Petroleum, Thessaloniki 221 000Greece Motor Oil Hellas, Corinth 1 440 000Greece Petrola Hellas, Elefsis 458 000Ireland Irish Refining Company, Whitegate 355 000 338 015



15 and Norw


61


[ton CO2 in 2001]



Comments

Italy ENI S.p.A Taranto 837 000Italy ENI S.p.A Livorno 561 000Italy IES Italiana Energia e Servizi, Mantova 364 000Italy Iplom Busallla 244 000Italy Raffineria di Augusta 2 240 000Italy Raffineria Falconara Marittima 1 510 000Italy Raffineria di Gela 3 610 000Italy Raffineria di Milazzo, Messina 2 320 000Italy Raffineria di Roma 385 000Italy Sannazzaro, Agip 1 950 000Italy Saras, Sarroch 5 990 000Italy Sarpom, Trecate 1 310 000Italy ENI S.p.A Port Maghera (Raffineria di

Venezia)703 000

Italy Priolo Siracusa, Praoil 1 680 000Italy Priolo Gargallo, Isab SPA 2 640 000Italy La Spezia < threshold No value of CO2 intensity could

be calculatedItaly Tamoil, Cremona 493 000Netherlands Esso Nederland / Raffinaderij Rotterdam 2 160 000 2 658 806Netherlands Q8 Europort 491 000 618 744Netherlands NEREFCO Europoort 2 240 000 2 239 380Netherlands Shell Nederland Pernis Rotterdam 6 300 000 6 580 258Netherlands Esha Smid and Hollander Amsterdam Not in EPER The refinery is opted out due to

emissions < 25 kton.Netherlands Total Raffinaderij Nedeland BV, Vlissingen 1 350 000 1 689 940Norway Esso Norge AS, Slagen 352 000Norway Statiol, Mongstad 1 530 000Portugal Sines 1 440 000Portugal Porto 958 000



15 and Norw


62


[ton CO2 in 2001]



Comments

Spain BP Oil castellon 745 000Spain Petroleos del Norte 2 060 000Spain Raffineria di Gibraltar (Cadiz) 1 790 000Spain Raffineria la Rabida 942 000Spain Raffineria Tenerife 459 000Spain Repsol Puertollano 2 850 000Spain Repsol Cartagena 814 000Spain Repsol la Coruna 1 580 000Spain Repsol Tarragona 2 710 000Sweden Preem Raff Göteborg 534 000 502 226Sweden Shell Göteborg 573 000 576 245Sweden Scanraff Lysekil 1 100 000 1 123 425UK BP Oil Coryton Refinery 2 070 000 2 255 071UK BP Oil Grangemouth Refinery Ltd 2 310 000 1 616 960UK Conoco Ltd Humber Refinery 2 010 000 2 639 008UK Eastham Refinery < threshold Not included in surveyUK Elf Oil UK Milfordhaven 1 060 000 1 175 557UK Esso Refinery Fawley 2 740 000 3 796 970UK Lindsey Oil Refinery 1 940 000 2 418 272UK Nynas UK Dundee < threshold Not included in surveyUK Petroplus Refining Teessside 272 000 300 226UK Shell UK Ltd Stanlow Manufacturing (Wirral) 2 800 000 2 926 576UK Texaco Pembroke 2 980 000 2 033 647



15 and Norw


63

IVL Svenska Miljöinstitutet AB

IVL Swedish Environmental Research Institute Ltd

P.O. Box 210 60, SE-100 31 Stockholm P.O. Box 5302, SE-400 14 GöteborgVisit: Hälsingegatan 43, Stockholm Visit: Aschebergssgatan 44Tel:: +46 (0)8 598 563 00 Tel:: +46 (0)31 725 62 00Fax: +46 (0) 8 598 563 90 Fax: +46 (0)31 725 62 90

www.ivl.se

Allowance Allocation and CO2 intensity of the EU15 and ......Allowance Allocation and CO2 intensity...

Documents

Transcript of Allowance Allocation and CO2 intensity of the EU15 and ......Allowance Allocation and CO2 intensity...